Masks for prevention of viral respiratory infections among health care workers and the public: PEER umbrella systematic review

Nicolas Dugré; Joey Ton; Danielle Perry; Scott Garrison; Jamie Falk; James McCormack; Samantha Moe; Christina S. Korownyk; Adrienne J. Lindblad; Michael R. Kolber; Betsy Thomas; Anthony Train; G. Michael Allan

Article Figures & Data

Figures

Tables
Additional Files

Download figure
Open in new tab
Figure 1.
Daily mask use compared with no mask use in the community to prevent influenzalike illness
^*Mantel-Haenszel fixed-effects method.

View popup

Table 2.

Study characteristics of mask use in health care workers to prevent viral respiratory tract infections

RCT (CLUSTER: YES OR NO)	COUNTRY,SETTING	POPULATION (AGE)	SAMPLE SIZE ENROLLED (ANALYZED)	CLUSTERS (RANDOMIZED)	INTERVENTIONS	WHOWORETHEMASKS	MASK USERECOMMENDATION	ADHERENCE
Jacobs et al,²⁸ 2009 (no)	Japan, tertiary care hospital	Health care workers (mean 35.5 y)	33 (32)	NA	Mask (17), no mask (15)	Health care workers (not sick)	Health care workers wear masks while on hospital property and performing their roles	84.3% of participants self-reported “full compliance,” with the remaining complying 79% to 99% of the time (applies to both mask use and nonuse)
Loeb et al,²⁹ 2009 (no)	Canada, tertiary hospitals	Nurses (mean 36 y)	446 (422)	NA	Surgical masks (212), N95 masks (210)	Nurses	When caring for patients with febrile respiratory illness	All 11 participants allocated to surgical masks wore them when caring for patientsadmitted to unit in droplet precautions for influenza
MacIntyre et al,³⁰ 2011 (yes)	China, hospital (ED and respiratory wards)	Nurses, doctors, ward clerks (mean 34 y)	1441 (1441)	Cluster: unit of randomization was hospital (15 hospitals involved, 5 per study arm)	Surgical masks (492), N95 fit-tested masks (461), N95 masks not fit-tested (488)	Health care workers	Every shift (given 3 surgical masks daily or 2 N95 masks daily)	Worn >80% on working days: surgical, 76% (5 h/day); N95 fit, 74% (5.2 h/day); N95 no fit, 68% (4.9 h/day)
MacIntyre et al,³¹ 2013 (yes)	China, hospitals (ED and respiratory wards)	Nurses, doctors (mean 33.1 y)	Surgical mask = 572 (572), targeted use of N95 = 561 (516), N95 = 581 (581), total = 1669 (166)	68 wards at 19 sites	Surgical masks (572), targeted use of N95 masks (516), N95 masks (581)	Nurses, doctors	All the time (surgical masks), as needed (targeted N95 masks), all the time (N95 masks)	66% for surgical masks, 82% for N95 targeted masks, 57% for N95 masks
MacIntyre et al,³² 2015 (yes)	Vietnam, hospitals (ED, ICU, ID or respiratory wards, pediatric ward)	Nurse or doctor (mean 36 y)	1607 (1607)	74 wards at 14 sites	Surgical masks (580), cloth masks (2-layer cotton) (569), control (458)—“standard practice” of mask use	Nurses or doctors	Surgical masks: all the time on shift; cloth masks: all the time on shift; control: standard practice	Surgical masks 56.6%, cloth masks 56.8%, standard practice 23.6%
Radonovich et al,³³ 2019 (yes)	US, outpatient sites (clinics, primary care clinics, EDs)	Health care personnel (mean 43 y)	5180 health care personnel seasons = 4051 participants (5180 seasons)	380 clusters at 137 sites	Surgical masks (2668 person-seasons), N95 masks (2512 person-seasons)—note that 1 person could be in different arms each of the 4 seasons	Those involved in direct patient care	Whenever positioned within 6 ft of a patient with suspected or confirmed respiratory illness	N95: 65.2% always, 24.2% sometimes; surgical: 65.1% always, 25.1% sometimes

ED—emergency department, ICU—intensive care unit, ID—infectious disease, NA—not applicable, RCT—randomized controlled trial, US—United States.

View popup

Table 3.

Outcomes for mask use to prevent viral respiratory tract infections

MASK USERS	CONFIRMED INFLUENZA		CONFIRMED VIRAL		INFLUENZALIKE ILLNESS		ANY RESPIRATORY INFECTION

	*RCTS (N^)**	RR^† (95% CI)	*RCTS (N^)**	RR^† (95% CI)	*RCTS (N^)**	RR^† (95% CI)	*RCTS (N^)**	RR^† (95% CI)
Community
• Community members	2 (1683)	0.93 (0.50–1.75)	0 (0)	NA	2 (1683)	0.83 (0.69–0.99)	0 (0)	NA
• Families—sick wearing masks	0 (0)	NA	1 (597)	0.98 (0.06–15.54)	2 (794)	0.89 (0.48–1.66)	1 (597)	0.65 (0.19–2.28)
• Families—healthy wearing masks	1^‡ (220)	2.18 (0.25-19.16)	1^‡ (220)	2.96 (0.67–13.02)	1‡ (220)	1.17 (0.62–2.18)	0 (0)	NA
• Families—healthy and sick wearing masks	4 (1168)	0.95 (0.71–1.26)	0 (0)	NA	4 (1377)	1.09 (0.86–1.37)	0 (0)	NA
• Communities staying in tents	1 (2317)	1.24 (0.33–4.59)	2 (2386)	1.18 (0.79–1.78)	1 (69)	0.56 (0.29–1.08)	1 (1810)	1.08 (0.83–1.40)
• Families—given masks before illness	0 (0)	NA	0 (0)	NA	1 (789)	0.90 (0.60–1.34)	0 (0)	NA
Health care workers
• Masks vs nothing	1 (431)	Not estimable	1 (431)	0.90 (0.33–2.44)	1 (431)	0.26 (0.01–6.42)	2 (463)	0.74 (0.36–1.54)
• N95 masks vs surgical masks	4^§ (6039)	1.10 (0.91–1.32)	4^§ (6059)	0.95 (0.83–1.07)	4^§ (6662)	0.78 (0.61–1.00)	3^§ (5033)	0.95 (0.90–1.00)
• Surgical masks vs cloth masks	1 (717)	0.98 (0.06–15.62)	1 (717)	0.62 (0.31–1.26)	1 (717)	0.12 (0.02–0.98)	1 (717)	0.62 (0.34–1.11)

NA—not applicable, RCT—randomized controlled trial, RR—risk ratio.
↵* The number of patients is modified (reduced) to account for clustering.
↵† Results calculated using a fixed-effects model.
↵‡ One RCT had 2 arms.
↵§ Two of the included RCTs had 2 arms, so 4 RCTs would be 6 arms and 3 RCTs would be 5 arms.