Background
-
To identify and categorise quality improvement approaches for small/sick hospitalised newborns in LMICs
-
To identify and categorise outcomes investigated by quality improvement initiatives for small/sick hospitalised newborns in LMICs
-
To identify barriers and promoters, at a local level and systems level, to the implementation of quality improvement initiatives for small/sick hospitalised newborns in LMICs
Methods
Eligibility criteria
-
Populations: hospitalised small and/or sick newborns in LMICs and admitted for inpatient healthcare. LMICs were identified according to the World Bank list of LMICs [26]. Facilities for this population must be defined as ‘hospitals’ or units within hospitals.
-
Interventions: quality improvement initiatives, according to the Ovretveit definition—“better patient experience and outcomes achieved through changing provider behaviour and organisation through using a systematic change method and strategies” [9].
-
Outcomes: objective clinical outcomes relating to mortality, morbidity, and process of care measures.
-
Language: studies published in English, or with translation available.
-
Year: published from 2000 or later.
Information sources
Search strategy
-
Term 1: Quality improvement(Quality or performance or effectiveness) AND (care or improvement* or increase* or service$ or indicator$)
-
Term 2: NewbornsNeonat* or neo-nat* or Baby or Babies or Newborn$ or new-born$ or infant$
-
Term 3: HospitalisedInpatient$ or in-patient$ or hospitalis* or NICU or neonatal intensive care unit.These search terms were then combined to give a final search of Term 1 AND Term 2 AND Term 3, which was used to search abstracts in these databases.
Study selection, extraction, and analysis
Results
Study selection
Study and participant characteristics
Author | Study type | Location | Population | Sample size | QI measure | Outcomes |
---|---|---|---|---|---|---|
Agarwal et al. 2007 [36] | Controlled before and after study | India | Neonates born within the obstetric teaching hospital | 15,249 | Package of interventions including rational practice, protocol usage, training and empowerment of nurses | Mortality: 30% decline in NMR Length of admission: reduction from 8.6 days to 7.1. Sepsis: reduction in deaths due to sepsis from 37.9 to 15.5% Appropriate antibiotic use: antibiotics use decreased to 23.2%. |
Bastani et al. 2015 [48] | Randomised controlled trial | Iran | Mothers with preterm infants | 91 | A family centred care (FCC) programme | Hospital admission/readmission: FCC group were significantly less likely to be rehospitalised, p = 0.04. Length of admission: 6.96 in FCC group, 12.96 in control group, p < 0.001. Maternal satisfaction: FCC group were significantly more satisfied. |
Bhutta et al. 2004 [49] | Controlled before and after study | Pakistan | Very low birth weight infants | 509 | A step-down unit for mothers and babies | Mortality: rates of survival increased, from 65 to 84% (p < 0.05). Length of admission: length of stay fell from mean of 34 to 16. Patient weight gain: there was a reduction in mean weight at discharge from 1.6 to 1.289 kg (p < 0.001). Patient infection rates: rates of overall nosocomial infections dropped significantly. Sepsis: increased relative risk of culture proved neonatal sepsis (95% CI 0.92–1.26). Severe illness: non-significant changes in rates of intraventricular haemorrhage, apnoeic spells, respiratory distress, and necrotising enterocolitis; significant decrease in patent ductus arteriosus rates. Presence of hyperbilirubinaemia: rates fell from 28.8% to 17.9%. |
Cavicchiolo et al. 2016 [58] | Controlled before and after study | Mozambique | NICU residents—inborn and outborn patients of all gestational ages up to the postnatal age of 7 days | 4276 | A continuous multi-level quality improvement intervention focused on infrastructure, equipment and protocol refinement | Mortality: reduction in death rate from 26 to 18%, significant. Hospital admission/readmission rate: admissions for prematurity, sepsis and asphyxia increased significantly. Sepsis: admissions for sepsis increased significantly, deaths decreased non-significantly. Severe illness: deaths for asphyxia increased significantly, admissions increased significantly. |
Clark et al. 2012 [50] | Controlled before and after study | Sierra Leone | Children presenting for emergency care | 500 | Training course based on ETAT WHO course, ward combined to form ICU and ER, triage area created, improved equipment, experienced nurses in triage, structured clerking pack introduced | Mortality: decreased from 12.38 to 5.85%. Length of admission: no change. |
Crouse et al. 2016 [38] | Controlled before and after study | Guatemala | Random sample of all patients presenting to the PED and all patients admitted to the PICU | 1027 | Emergency Triage Assessment and Treatment (ETAT)-based emergency triage process | Mortality: decreased from 12 to 6% amongst critically ill, not significant. Hospital admission/readmission: admission from the Paediatric Emergency Department fell significantly from 8 to 4%, and also fell significantly in critically ill group. Length of admission: decreased, not significant. |
Darmstadt et al. 2005 [51] | Controlled before and after study | Bangladesh | Preterm infants in Special Care Nursery | – | Infection control programme | Mortality: decline in deaths of certain causes, significance not mentioned. Patient infection rates: decrease in nosocomial infection reports, and K. pneumoniae. Infection detection rates: decline in cases of culture-proven sepsis and suspected sepsis. Sepsis: significant decline in patients with clinical diagnosis of sepsis (79%). Appropriate antibiotic use: antibiotic use guidelines were reviewed, no data. Adherence to national guidelines of care: staff trained in standard guidelines, antibiotic guidelines were adhered to. |
dos Santos et al. 2015. [47] | Intervention study (non-random) | Brazil | NICU newborns | 24 | NIPS scale; non-pharmacological actions in pain control in newborns | Adherence to national guidelines of care: significantly lower NIPS (pain scale) score with intervention. |
Erdeve et al. 2008 [52] | Intervention study (non-random) | Turkey | All mother−preterm infant dyads that were consecutively admitted to the NICU | 60 | Use of individual rooms | Hospital admission/readmission: rehospitalisation rate was higher in non-intervention group p < 0.05. Length of admission: no significant difference regarding duration of intensive care hospitalisation. Patient weight gain: no significant change on discharge in body weight. Breastfeeding practice: no significant change in groups regarding breastfeeding rates. |
Gathara et al. [39] | Controlled before and after study | Kenya | Sick newborns aged 0–7 days and malnourished children aged 6–59 months | 798 | Package of interventions including clinical guidance booklets, admission record form, a training course on emergency and admission care, external support supervision, local facilitation, performance assessment, and feedback | Mortality: mortality was reduced by 3% post intervention in intervention group, control group was static. Appropriate antibiotic use: overdoses of penicillin were reduced in intervention vs control group, but overdoses of gentamicin were increased. Adherence to national guidelines of care: documentation of gestation in weeks were increased in intervention group, and mean documentation score was higher. More vitamin K was prescribed in intervention groups. |
Gilbert et al. 2014 [53] | Controlled before and after study | Brazil | Neonates admitted to NICU | 1242 | A 5-phase POINTS of Care package | Mortality: crude survival rates did not change over time significantly except in one NICU where it decreased. Patient weight gain: days to regain birth weight were significantly higher in post-intervention period. Retinopathy of prematurity: no significant change. Sepsis: rates did not change—11.3/12.3 cases per 1000 infant days. Lower respiratory tract disease: non-significant increase in bronchopulmonary dysplasia. Severe illness: non-significant increase in bronchopulmonary dysplasia, no change in necrotising enterocolitis. |
Leng et al. 2016 [40] | Controlled before and after study | China | Very low birth weight neonates | 172 | Use of radiant warmers, warmer delivery room, STABLE programme, consulting services, standardised transportation, education of staff, review and feedback | Mortality: mortality rates decreased from 12 to 7%, p = 0.03. Length of admission: reduced from 60 to 45 days, p = 0.01 Sepsis: sepsis rates did not change significantly. Hypothermia rates: significant decrease in patients with temperatures < 36 degrees Celsius. Lower respiratory tract disease: percentage with chronic lung disease did not change significantly. Severe illness: rates of intraventricular haemorrhage and necrotising enterocolitis did not change significantly, but SNAPPE-II score increased significantly. |
Mais et al. 2015 [41] | Controlled before and after study | Lebanon | Neonates with central lines in NICU | 213 | Theoretical and practical teaching sessions, dressing change guidelines, sterile technique, auditing adherence to guidelines | Length of admission: there was no significant change. Patient infection rates: CLABSI rates declined significantly, p < 0.05. Mechanical ventilation: no significant change. Central line duration: no significant decline in usage. |
Namazzi et al. 2015 [42] | Controlled before and after study | Uganda | All pregnant and newly delivered mothers residing within the villages of the Iganga/Mayuge Health and Demographic Surveillance Site | – | District led training, support supervision, mentoring, supply of essential medicine and equipment | Mortality: hospitalised NMR declined from 17 to 9%, not significant. Kangaroo Mother Care: by the end of the study, 547 preterm babies had been cared for in a KMC unit. Premature delivery rate: rate was 8% in deliveries in health units. |
Pinto et al. 2013 [43] | Controlled before and after study | Brazil | Newborns with very low birth weight | 136 | Dissemination of a new protocol proposed by the Brazilian National Health Surveillance Agency for antibiotic usage in LBW infants | Mortality: overall mortality decreased from 20.9 to 4.4%, significant. Patient infection rates: no significant change in multi-resistant infection rates. Sepsis: no difference in relation to confirmed sepsis, but a significant reduction in diagnoses of probable sepsis. Severe illness: no change in diagnoses of severe illnesses, e.g., PDA, PBD, necrotising enterocolitis. Appropriate antibiotic use: decrease in number of antimicrobial regimens used and days of antibiotic use. |
Rahman et al. 2017 [44] | Controlled before and after study | Bangladesh | Children identified as having systemic sepsis | 1036 | Triage, fast assessment, immediate results, immediate antibiotics, training package, slow charts, checklist, records system, infection control measures, equipment stocking | Mortality: mortality decreased, significance not reported. Length of admission: increase in % with syndromic sepsis staying for over 48 h, significance not reported. Appropriate oxygen use: post intervention 94% were given oxygen with hypoxaemia. Appropriate antibiotic use: first-line recommended antibiotic usage increased from 49 to 75%, p < 0.005. |
Ramaswamy et al. 2015 [59] | Controlled before and after study | Ghana | Obstetric and neonatal cases in regional referral facilities | – | Ridge-Kybele model for obstetric and neonatal care—an integrated approach to systems change | Adherence to national guidelines of care: 37% improvement in NICU hand hygiene rates. Waiting times: 74% reduction in mothers with unacceptable waiting times. |
Rosenthal et al. 2012 [33] | Controlled before and after study | Argentina, Colombia, El Salvador, India, Mexico, Morocco, Peru, the Philippines, Tunisia, Turkey | NICU patients | 6829 | VAP (ventilator-associated pneumonia) bundle—11 items | Patient infection rates: ventilator-associated pneumonia rates per 1000 mechanical ventilator days decreased from 17.8 to 12.0. Lower respiratory tract disease: ventilator-associated pneumonia rates per 1000 mechanical ventilator days decreased from 17.8 to 12.0. Adherence to National Guidelines of Care: hand hygiene compliance rates rose from 62 to 81%. Mechanical ventilation: days of MV did not change. |
Rosenthal et al. 2013 [34] | Controlled before and after study | El Salvador, Mexico, Philippines, and Tunisia | NICU patients with central line insertion | 2214 | INICC multidimensional infection control approach | Patient infection rates: CLABSI rate reduction from baseline of 54%, 95% CI 0.33–0.63 RR. Adherence to National Guidelines of Care: hand hygiene and sterile gauze rates rose significantly. |
Salehi et al. 2015 [45] | Controlled before and after study | Iran | Hospitalised ‘infants’ | 100 | Implementation of guidelines and education | Patient weight gain: patients in intervention group had a mean weight change of + 96 g compared to − 59, p = 0.001. |
Sethi et al. 2017 [54] | Controlled before and after study | India | Preterm neonates | 26 neonates, 23 mothers | CPNC—comprehensive post-natal counselling package, comprising education of health care providers and family members | Breastfeeding practice: the proportion of mothers expressing milk on day 1 increased to 86.6% from 12.5%, after 1 year the proportion of neonates on exclusive breast milk was more than 80%. |
Soni et al. 2016 [55] | Controlled before and after study | India | Infants admitted to a rural Indian neonatal intensive care unit (NICU) | 648 | Presence of physician champions | Length of admission: length of stay was greater with champions, at 9 days, compared to 7 without, p = 0.01. Patient infection rates: patients who experienced infections decreased significantly as physician champions left. Appropriate antibiotic use: no association between champions and antibiotic usage. Breastfeeding practice: breastfeeding rates were not changed. Usage of Kangaroo Mother Care: skin to skin care increased with champions and lasted longer hours per day. Premature delivery rate: with KMC champions there was a higher percentage of premature deliveries, p = 0.01 for trend. |
Srofenyoh et al. 2012 [35] | Controlled before and after study | Ghana | Mothers and neonates in Ridge Regional Hospital | 29,508 | An interdisciplinary approach, high-level sponsorship, establishment of guidelines, measurement, feedback, leadership and teamwork coaching, training including QI training, and a multimodal focus on patients, providers, and systems | Mortality: perinatal mortality was reduced, no information on significance. Maternal satisfaction: this improved. Maternal health: 34% decrease in maternal mortality. Stillbirth: reduced by 36%, p < 0.05. |
UNICEF 2014 [37] | Controlled before and after study | Bangladesh | Hospitalised newborns | – | Quality improvement initiatives delivered alongside SCANUs—Special Care Newborn Units | Mortality: average case fatality rates dropped in most SCANUs. Hospital admission/readmission: admissions at SCANUs increased. |
Wrammert et al. 2017 [56] | Controlled before and after study | Nepal | Neonates in maternity hospital, Kathmandu | 299 | Implementation of Helping Babies Breathe Protocol | Mortality: decrease in death rate in first 24 h, p < 0.01. No significant change in 7/28 day mortality. |
Yawson et al. 2016 [60] | Controlled before and after study | Ghana | Users of Ghanaian newborn care service | – | BNA tool to identify service gaps with group discussions, leading to national and regional operational plans and monitoring/evaluation framework | Mortality: mortality reduced in the intervention regions. |
Zhou et al. 2013 [57] | Controlled before and after study | China | All neonates who received mechanical ventilation for at least 48 h and were hospitalised in the NICU for ≥ 5 days | 491 | A bundle of comprehensive preventive measures against VAP were gradually implemented using the evidence-based practice for improving quality method. | Mortality: mortality rates decreased from 14% in phase 1 to 3% in phases 2 and 3, statistically significant. Patient infection rates: sustained decline in VAP rates, p = 0.01. |
Zhou et al. 2015 [46] | Controlled before and after study | China | Neonates in the NICU | 171 | EPIQ programme—team taught for 2 days, who then identified strategies for adoption of CLABSI prevention, and trained other members | Patient infection rates: CLABSI rates declined in each successive phase. Central line duration: time in situ increased across the phases, significance not reported. |
ROBINS-I tool for non-randomised studies of interventions | ||||||||
Studies | Bias due to confounding | Bias in selection of participants into the study | Bias in classification of interventions | Bias due to derivations from intended interventions | Bias due to missing data | Bias in measurement of outcomes | Bias in selection of the reported result | Overall bias |
Agarwal et al. [36] | Moderate | Low | Low | Low | Low | Low | Moderate | Moderate |
Bhutta et al. [49] | Moderate | Low | Low | Low | Low | Moderate | Serious | Serious |
Cavicchiolo et al. [58] | Moderate | Low | Low | Low | Low | Moderate | Moderate | Moderate |
Clark et al. [50] | Serious | Low | Low | Low | Low | Low | Moderate | Serious |
Crouse et al. [38] | Moderate | Low | Low | Low | Serious | Low | Moderate | Serious |
Darmstadt et al. [51] | Moderate | Low | Low | Low | No info | Low | Moderate | Moderate |
Dos Santos et al. [47] | Serious | NI | Serious | Low | Low | Serious | Moderate | Serious |
Erdeve et al. [52] | Moderate | Low | Low | Low | Low | Moderate | Moderate | Moderate |
Gathara et al. [39] | Moderate | Moderate | Low | Low | Serious | Moderate | Moderate | Serious |
Gilbert et al. [53] | Moderate | Low | Low | Low | Moderate | Low | Moderate | Moderate |
Leng et al. [40] | Moderate | Low | Low | Low | Low | Moderate | Moderate | Moderate |
Mais et al. [41] | Moderate | Low | Low | Low | Low | Low | Moderate | Moderate |
Namazzi et al. [42] | Serious | Low | Low | Low | No info | Moderate | Moderate | Serious |
Pinto et al. [43] | Moderate | Low | Low | Low | Low | Low | Moderate | Moderate |
Rahman et al. [44] | Moderate | Low | Low | Low | No info | No info | Moderate | Moderate |
Ramaswamy et al. [59] | No info | No info | No info | No info | No info | No info | Serious | Serious |
Rosenthal et al. [33] | Moderate | Low | Low | Low | Low | Moderate | Moderate | Moderate |
Rosenthal et al. [34] | Moderate | Low | Low | Low | Low | Low | Moderate | Moderate |
Salehi et al. [45] | Serious | Low | Low | Low | No info | Low | Low | Serious |
Sethi et al. [54] | Serious | Low | Low | Low | Low | Low | Moderate | Serious |
Soni et al. [55] | Moderate | Low | Low | Low | Serious | Low | Moderate | Serious |
Srofenyoh et al. [35] | Moderate | Low | Low | Low | Low | Low | Moderate | Moderate |
UNICEF [37] | No info | No info | No info | No info | No info | No info | Critical | Critical |
Wrammert et al. [56] | Moderate | Low | Low | Low | Low | Low | Moderate | Moderate |
Yawson et al. [60] | No info | No info | Low | Low | No info | No info | Serious | Serious |
Zhou et al. [57] | Serious | Low | Low | Low | Low | Low | Moderate | Serious |
Zhou et al. [46] | Serious | Low | Low | Low | Low | Low | Moderate | Serious |
Risk of Bias 2.0 tool for randomised studies | ||||||||
Studies | Bias arising from the randomisation process | Bias due to deviations from intended interventions | Bias due to missing outcome data | Bias in the measurement of the outcome | Bias in the selection of the reported result | Overall bias | ||
Bastani et al. [48] | Low | Medium | Low | Medium | Low | Medium |
Classified quality improvement approaches
Level | Strategy | Total | Citation |
---|---|---|---|
Micro | Distribution of referencing materials to providers | 8 studies | |
Decision support | 2 studies | ||
Care coordination | 5 studies | ||
Meso | Strengthening facility infrastructure | 6 studies | |
Continuous quality improvement | 7 studies | ||
Supervision | 5 studies | ||
Feedback | 6 studies | ||
In-service training | 20 studies | ||
Service organisation | 9 studies | ||
Macro | Regulation and governance | 1 study | [59] |
Task shifting | 1 study | [60] |
Groups of outcomes measured in quality improvement approaches
Quality of care classification of QI outcome measure | Quality improvement outcome | Significant increase | Significant decrease | No significant change | Significance not assessed or not reported |
---|---|---|---|---|---|
Safe (minimising risks and harm) | Mortality | – | |||
Patient weight gain | 1 study—[45] | 1 study—[52] | – | ||
Patient infection rates | 1 study—[55] | 1 study—[43] | – | ||
Effect on retinopathy of prematurity | – | – | 1 study—[53] | – | |
Sepsis rates | – | – | |||
Rates of hypothermia | – | 1 study—[40] | – | – | |
Patient lower respiratory tract disease | – | 1 study—[33] | – | ||
Severe illness (various) | 1 study—[49] | – | |||
Presence of hyperbilirubinaemia | – | – | 1 study—[49] | – | |
Effect on breastfeeding practice | 1 study—[54] | – | – | ||
Maternal health | – | 1 study—[35] | – | – | |
Stillbirth | – | 1 study—[35] | |||
Premature delivery rate | 1 study—[55] | – | – | 1 study—[42] | |
Effective (utilising evidence) | Appropriate oxygen use | – | – | – | 1 study—[44] |
Antibiotic usage | 1 study—[44] | 1 study—[55] | |||
Adherence to national guidelines of care | – | – | |||
Mechanical ventilator days | – | – | – | ||
Central line duration | – | – | 1 study—[41] | 1 study—[46] | |
Efficient (avoiding waste) | Length of admission | 1 study—[55] | |||
Hospital admission/readmission | 1 study—[58] | – | 1 study—[49] | ||
People-centred (accounting for preferences of service users) | Usage of Kangaroo Mother Care | 1 study—[55] | – | – | 1 study—[42] |
Maternal satisfaction | 1 study—[48] | – | – | 1 study—[35] | |
Timely (reducing delays) | Waiting times | – | – | – | 1 study—[59] |
Mortality rate
Length of admission
Sepsis rates
Patient infection rates
Barriers and promoters to implementing quality improvement approaches
Local level | Studies | Systems level | Studies |
---|---|---|---|
Promoters | |||
Motivation of key individuals | Relationships between health workers, community leaders and district officials | 1 study—[42] | |
Continuous monitoring throughout | High-quality national data collection | 1 study—[60] | |
Interdisciplinary collaboration | Formal health service support | 1 study—[35] | |
Abandonment of unnecessary practices | 1 study—[36] | NGO collaboration initiatives | 1 study—[58] |
Schemes tailored to participants | 1 study—[38] | ||
On-site support | 1 study—[44] | ||
Refresher programmes | 1 study—[44] | ||
Formal training in QI methods | 1 study—[35] | ||
Low cost of intervention | 1 study—[38] | ||
Barriers | |||
Overburdened staff | Insufficient funding | 1 study—[42] | |
Lack of sufficient equipment | Insufficient health services relative to demand | 1 study—[42] | |
High changeover of workforce | Government redistribution of staff | 1 study—[53] | |
Defects in staff knowledge and practice | 1 study—[35] | Inadequate documentation | 1 study—[39] |
Unmotivated staff | 1 study—[53] | Confounding health policy changes | 1 study—[50] |
Multiple QI measures/audits simultaneously | 1 study—[55] |