Table 4.

Characteristics of included studies, anaesthetic task-shifting

Study name and country	Study design	Healthcare facility characteristics	Procedure characteristics	Lead anaesthetic providers and number of procedures	Training and level of supervision of non-specialists	Outcomes assessed	Mortality	Other clinical outcomes
Cometto et al., 2012. South Sudan (sub-Saharan Africa). Low-income economy.	Retrospective review. Data were collected between 2005 and 2010.	Multi-centre study. The programme was implemented in Primary Healthcare Centres, supported by the Italian NGO Comitato Collaborazione Medica.	Anaesthetic task-shifting. 1543 patients received anaesthesia during the surgical missions of Comitato Collaborazione Medica. Most operations (60%) were performed under spinal anaesthesia. A minority of cases required ketamine anaesthesia with intravenous supplementation of analgesic drugs. Many cases of minor surgery were performed under local infiltration. Few cases (e.g. goitres and acute abdomen) were operated under general anaesthesia with endotracheal intubation and ventilation with Ambu balloon.	Anaesthesia technicians: non-physicians. They administered anaesthesia to 511 surgical patients. Visiting expatriate consultant surgeons: specialist physicians. They administered anaesthesia to 1032 surgical patients.	Anaesthesia technicians had a level of background training between nurses and physicians. They were trained through the training programme of Comitato Collaborazione Medica and through the War Wounded Referral Hospital (Kenya) managed by the International Committee of the Red Cross. They developed competencies in perioperative surgical care.	Mortality.	No differences between anaesthesia technicians and expatriate surgeons.
Dulisse and Cromwell, 2010. USA. High-income economy.	Retrospective review. Data were collected between 1999 and 2005 through the ‘Medicare Parts A and B claims’ limited datasets across ‘non-opt-out’ states and 14 ‘opt-out’ states. Opt-out states were those that allowed the reimbursement of certified registered nurse anaesthetists (CRNAs) operating independently through Medicare and Medicaid Services.	Multi-centre study.	Anaesthetic task-shifting. Anaesthesia was administered as part of surgical procedures.	CRNAs practicing independently (CRNA solo): non-physicians. CRNAs independently provided anaesthesia in 21% of surgeries in opt-out states and in 9.7% of surgeries in non-opt-out states between 1999 and 2005. Medical Doctor Anaesthesiologists (MDA) practicing independently : specialist physicians. They independently provided anaesthesia in 42% of surgeries in opt-out states and in 44.5% of surgeries in non-opt-out states between 1999 and 2005. Team (MDAs and CRNAs): teams provided anaesthesia in 37% of surgeries in opt-out states and 45.8% of surgeries in non-opt-out states between 1999 and 2005.	The training of CRNAs was not described in detail.	In-patient mortality and anaesthesia complications. Outcomes were calculated taking the ‘MDA solo’ group in non-opt-out states as the reference category. To evaluate task-shifting, authors focused on the comparisons between the ‘MDA solo’ group in non-opt-out states group and ‘CRNA solo’ groups (in non-opt-out and opt-out states).	CRNA solo’ in non-opt-out states: OR 0.899 (P = 0.05). ‘CRNA solo’ in opt-out states, before opting out: OR 0.651 (P = 0.05). ‘CRNA solo’ in opt-out States, after opting out: OR 0.689 (P = 0.05).	CRNA solo’ in non-opt-out states: OR 0.992 (not significant). ‘CRNA solo’ in opt-out states, before opting out: OR 0.798 (P ≤ 0.05). ‘CRNA solo’ in opt-out states, after opting out: OR 0.813 (P ≤ 0.05).
Kudsk-Iversen et al., 2020. 23 countries across the Africa Region, the Americas Region, the Eastern Mediterranean Region and the South-East Asia Region.	Retrospective observational study. Data were collected between January 2008 and December 2017.	Multi-centre study. Data were collected for 173 084 cases across 23 countries and 52 different locations. 28 surgical projects were set in armed conflict settings and 32 projects were in healthcare gap (HG) settings.	Anaesthetic task-shifting. Anaesthesia was administered as part of surgical projects delivered by Médecins Sans Frontières (MSF). Most common anaesthesia: spinal injection alone and general anaesthesia without intubation (mostly ketamine based).	MSF projects were classified based on the most senior anaesthetic provider. Uncertified anaesthesia providers: local non-physicians. They led 15% of cases in armed conflict settings and HG settings. Nurse anaesthetists: nurses or other non-physician clinical cadres who were predominantly from low-income settings. They led 19% of cases in armed conflict settings and HG settings. Anaesthesiologists: specialist physicians with qualifications in anaesthesia, either local or expatriates (from both low-income and high-income settings). They led 66% of cases in armed conflict settings and HG settings.	Uncertified anaesthesia providers had different levels of experience in anaesthesia provision, but they lacked formal qualifications. They received on-the-job training. Nurse anaesthetists had received formal training and qualification in anaesthesia in their country of origin.	Mortality (intraoperative).	Comparable mortality rates between provider groups. Armed conflict settings: 0.3% (uncertified providers), 0.2% (nurse anaesthetists) and 0.3% (anaesthesiologists). HG settings: 0.2% (uncertified providers), 0.1% (nurse anaesthetists) and 0.3% (anaesthesiologists).
Needleman and Minnick, 2009. USA (California, Florida, Kentucky, New York, Texas, Washington and Wisconsin). High-income economy.	Retrospective review. Data were collected in 1999–2001 (California, Florida, New York, Washington and Wisconsin) and 2000–2001 (Kentucky and Texas).	Multi-centre study across 369 hospitals. 27% of hospitals were located in California, 19% in Wisconsin, 14% in New York, 13% in Texas, 13% in Washington, 9% in Florida and 5% in Kentucky. 67% were metropolitan hospitals, 83% were non-teaching hospitals and 69% were non-profit hospitals.	Anaesthetic task-shifting. Anaesthesia was delivered for 1 141 641 obstetrical surgical procedures (caesarean section deliveries).	CRNAs only: non-physicians. Solely CRNAs administered anaesthesia in the study hospital. 23% of hospitals implemented a ‘CRNA-only’ model. Anaesthesiologists only: specialist physicians. Solely anaesthesiologists administered anaesthesia in the study hospitals. 39% of hospitals implemented an ‘anaesthesiologist-only’ model. Other models of anaesthesia provision included ‘ANES-CRNA I’, ‘ANES-CRNA II’ and ‘mixed’. To evaluate task-shifting, we focused on comparisons between the ‘anaesthesiologist-only’ and ‘CRNA-only’ groups.	The training of CRNAs was not described in detail.	Mortality (in-hospital), anaesthesia complications, anaesthesia or other complications (composite outcome) and obstetrical trauma (Agency for Healthcare Research and Quality Patient Safety indicators related to obstetrical trauma). Outcomes were calculated taking the ‘anaesthesiologist-only’ group as the reference category.	No significant differences between ‘CRNAs only’ and ‘anaesthesiologists only’ for mortality.	Anaesthesia or other complications, ‘CRNAs only’ vs ‘anaesthesiologists only’: OR 0.723 (0.542–0.965, P = 0.028). No significant differences between ‘CRNA-only’ and ‘anaesthesiologist-only’ groups for anaesthesia complications and obstetrical trauma.
van der Merwe et al., 2021. 25 countries across Africa.	Secondary analysis of the African Surgical Outcomes Study, a prospective observational cohort study. Data were collected over 7 days, between February and May 2016.	Multi-centre study. Patients treated by non-physicians: 25% in primary healthcare facilities; 27% in secondary healthcare facilities and 49% in tertiary healthcare facilities. 83% of hospitals were government funded and 17% were privately funded or mixed. Patients treated by physicians: 22% in primary healthcare facilities, 29% in secondary healthcare facilities and 50% in tertiary healthcare facilities. 85% of hospitals were government-funded and 15% were privately funded or mixed.	Anaesthetic task-shifting. Procedural sedation administered for adult elective and emergency in-patient surgical operations. General and regional anaesthesia were excluded. Severity, by provider type: non-physicians: 50% minor, 44% intermediate and 6% major. Physicians: 53% minor, 31% intermediate and 17% major. Urgency: non-physicians: 48% emergency. Physicians: 40% emergency.	Non-physicians administered anaesthesia to 98 patients. Physicians: specialists and non-specialists. They administered anaesthesia to 235 patients.	Training and level of supervision of non-physicians were not specified.	Severe complications and death (composite outcome).	12.8% (non-physicians) vs 1.6% (physicians), average treatment effect 8.3 (2.7–25.6), P < 0.001.	12.8% (non-physicians) vs 1.6% (physicians), average treatment effect 8.3 (2.7–25.6, P < 0.001).