Discussion
We found substantial variation in the numbers of ICU beds, as well as the capacity for mechanical ventilation and specialized support for respiratory failure among ICUs in Canada. These findings were not fully explained by the size of the population. This variation in capacity may result in differential decision-making about who can receive ICU support, and which services can be supported in specific hospitals and regions during times of increased demand [
3,
5].
Prior work by our group using health administrative data from the Canadian Institute for Health Information estimated that there were 319 ICUs, 3,388 total adult ICU beds (representing 3.4% of all acute care hospital beds), and 13.5 ICU beds per 100,000 population [
8]. However, these data were based on a more liberal definition of critical care beds, did not include data from Quebec, did not include any interprovincial comparisons, did not estimate the capacity to treat critically ill patients requiring mechanical ventilation, and were generated one decade ago. Our assessment of ICU beds per 100,000 population places Canada near the median of high-income and Organization for Economic Co-operation and Development countries, notably above the United Kingdom but well below the United States, Germany, and Belgium [
8,
7,
20,
21,
22].
Without knowledge of Canadian critical care capacity, and in the absence of provincial, national, or international targets for population-based critical care resources, there has been limited national attention to ensuring optimal distribution among regions. The results of this survey highlight expected north–south geography-based capacity trends, but also an unanticipated apparent east–west gradient with relative increased capacity among the Atlantic Provinces, in comparison with central and western Canada. Some members of our group have previously reported wide variation in ICU capacity within British Columbia and an inverse relationship between ICU beds, population density, and population growth, highlighting the potential for mismatch in demand and capacity in Canada [
11].
Some variation in distribution of healthcare services is probably a consequence of differential regional models of healthcare delivery. For example, there are substantial differences in population density across Canada, with marked north-to-south increases in density – approximately 80% of Canadians live within 160 km of the Canada–United States border [
23] – and there is marked intra-provincial urban–rural variability [
24]. Many systems of intra-provincial regionalization of care that are responsive to population density, geographic barriers, or evolved regional care systems also may lead to differences in the distribution of critical care services. Moreover, despite provincial administration of most healthcare services, there is also some well-established inter-provincial ICU care, with regionalized trauma services, specialized care delivery for northern territories among bordering provinces, and populations of one province that are closest to a specialized healthcare center in a neighboring province. It is not clear, however, whether specialized services’ referral relationships work well during times of healthcare crisis such a pandemic when regions may react to future uncertainty by trying to conserve resources for more local use.
This study has important limitations. First, this survey was carried out using existing national and provincial databases of hospitals, and it is possible that some acute care hospitals may have been missed. However, we subsequently employed snowball sampling and web and map searching techniques to identify all hospitals and ICUs in each province, and then sought out a combination of physician, nurse, respiratory therapist, and hospital administrator leaders to derive current ICU beds and ventilator capacity at each hospital. After compiling local data, each participant and provincial health authority was given the opportunity to critique the aggregate estimates to improve accuracy. Second, population denominator-based comparisons may not be the optimal mechanism for normalization in all regions with varying population density, age demographic differences, geographic barriers, and distinct systems of regionalized care for some tertiary and quaternary services such as trauma and transplantation. However, our results indicate relatively wide variability in ICU capacity among provinces and therefore may provide helpful inter-provincial comparisons. Third, this study focused on a very narrow spectrum of services needed to provide critical care – ICUs, beds, ventilators, and specialized supports for respiratory failure. It was beyond the scope of this survey to evaluate personnel (dieticians, nurses, pharmacists, physicians, physiotherapists, respiratory therapists, social workers) or other resources that are essential to the care of critically ill patients. Indeed, lack of available critical care clinical staff is among the most common reason for limitations in bed availability [
25-
27]. Future resource planning must address this key knowledge gap. Fourth, ICU resources are not static, and this survey represents a period prevalence of approximately 3 months at the hospital level and approximately 1 year among all sites, in a period after the H1N1 pandemic where knowledge of ICU capacity may have been greatest.
Our results highlight the need to examine capacity both in relation to local needs and in comparison with other regions. It is important to note that the organization of critical care within Canada has not been static since conducting this survey. Alberta has reorganized critical care services under one structure, with a standardized provincial bedside clinical information system/electronic medical record [
28]. Since the severe acute respiratory syndrome experience, the Ministry of Health and Long-term Care in Ontario has maintained a Critical Care Strategy to oversee a similar cataloguing of critical care services including twice-daily clinical updates of every patient in ICUs into a centralized electronic database that facilitates critical care inter-facility transportation services, reporting on quality metrics and decision-making on surge capacity [
29]. British Columbia and Nova Scotia have recently formed Critical Care Working Groups within the Ministry of Health to coordinate data collection and reporting, improvement of care processes, transportation of critically ill patients, and improvement of staffing models in ICUs. In 2011 Quebec created a Groupe d’Experts en Soins Intensifs working with the Ministry of Health to improve quality and accessibility of ICUs and a mandate to establish provincial ICU capacity.
One of the lessons learned from the severe acute respiratory syndrome and influenza A (H1N1) pandemics is that infectious outbreaks do not respect regional health boundaries [
30,
31] and that individual regions may be clinically overwhelmed while others are unaffected. Of relevance to surge planning, we were able to quantify the excess numbers of invasive mechanical ventilators relative to ICU beds, highlighting capacity that may exist beyond existing ICU beds. The ability to provide advanced oxygenation with one of three modes of support was available in a minority of hospitals. Furthermore, this expertise was unevenly distributed across provinces, and was focused at university-affiliated teaching hospitals. However, we were unable to gauge experience with specialized ventilation alongside capacity. We did not determine capacity for other techniques such as prone ventilation, which may be less dependent upon specific technology, more dependent upon generation of a local experience base, and have a greater evidence base for efficacy than either early high-frequency ventilation or use of inhaled nitric oxide [
32]. This variable and uneven distribution of expertise observed in this study demands that we evolve a system in which excess capacity in one region may aid another, either through safe transportation of patients or short-term movement of equipment or personnel to existing or temporary facilities [
33].
Acknowledgements
Yukon collaborators: KD Braden, Nora Tremblay, Bendan Hanley, Rowena Beckett, and Karin Heynen.
Northwest Territories collaborators: Ivan Russell, Kami Kandola, Kathie Pender, and Elaine Kelly.
Nunavut collaborator: W. Alexander (Sandy) MacDonald.
British Columbia collaborators: Tracey Miller, Lynn Tran, Lauren Tindall, Michelle Stuart, Kevin Carriere, Elizabeth Jolley, Gordon Krahn, Jim Fitzpatrick, Lynn Smiley, Coleen Hay, Ophelia Spencer, Darla Roberts, Christina Gordon, Barb Caldwell, Jo-Annne Rondeau, Karen L Watson, Morag Mochan, Joanne Cozac, Cheryl Scovill, Kathey Yeulet, Suzanne Johnston, Ray Taylor, Pat Tresierra, Rob Fingland, Christopher Gagnon, Monique Roy-Michaeli, Grzegorz Martinka, David Wensley, Peter Skippen, Sneeta Takhar, and Don Griesdale.
Alberta collaborators: Monique Roy-Michaeli, Ari Joffe, Elaine Gilfoyle, P Wickson, Catherine Ross, Chip Doig, Paul Boiteau, and David Zygun.
Saskatchewan collaborators: Susan Lyons, Anita Sagadahi, Laurie Albinet, Gayle Riendeau, Pam McKay, Ndrew McLetchie, Barbara Jiricka, Carol Gregoryk, Val Davies, Valerie Mann, Saqib Shahab, Patrick O’Byrne, Diane Larrivee, Sandy Bradford-Macalanda, and Alice Wong.
Manitoba collaborators: Kimberly Webster, Scott Cleghorn, Gregg Eschun, Allan Garland, Charlie Penner, Murray Kesselman, and Gordon Kasian.
Ontario collaborators: Carol Moran, Jeff Singh, Judith Van Huyse, Sonny Dhanani, Jamie Hutchison, John Muscedere, Stephen Lapinsky, Kusum Menon, Karen Choong, Tom Stewart, Margaret Herridge, Douglas Fraser, Karen Burns, and Niall Ferguson.
Quebec collaborators: Baqir Quizibash, Jean Philippe Garant, Lysanne Desaindes, Vanessa Dutil, Craig Baldry, Francois Lellouche, Pjil Roula, Ash Gursahaney, Arnold Kristoff, J-S Bilodeau, Mathieu Bernier, Jean-Luc Houde, Marisa Tucci, Marc-Andre Dugas, Dina Diana, Miriam Santschi, Denny Laporta, Dominique Piquette, Davinia Withington, Germain Poirier, Martin Legare, Jacques Lacroix, Stephane P. Ahern, Dr. François Lauzier, and Tina Doyle.
New Brunswick collaborators: Carolin Galvin, Rachel Mallais, Marchel Mallet, Kathy Kowalski, Claude Violette, Stephanie Perry, Colleen Reinsborough, John Mowat, Todd Lambert, Krista Chillington, Judy Melanson, Joanne Michaud-Young, Marc Pelletier, Krista Shillington, and Jean Bustard.
Prince Edward Island collaborators: Judy Adams, Sherry Harris, and Kay Kelly.
Nova Scotia collaborators: Anne McClair, Linda Rouleau, Dietrich Henzler, Rebecca Earle, Nadine Stevens, Maria Marshall, Jill Smith, Wendy Studley, Norah Doucet, Angela Foote, Dorothy McCaskil, Sharon McCarthy, Kim Thomasl Kate Mahon, Shauna Best, Chris Soder, and Rick Hall.
Newfoundland and Labrador collaborators: Jill Barter, Gerry McCain, Kim Aduau, Sharon Penny, and Valery Clarke.
Public Health Agency of Canada collaborators: Rachel Rodin, Corey Oliver, Ahalya Mahendra, Joy Pulickal, Marianna Ofner, and Kara Hayne.
Authors’ contributions
RAF, PJ, DJC, JCM, and AFT conceived of the study. RF and PJ participated in the design of the study. RAF, SMB, NB, DJC, PD, DF, RF, NG, RG, PJ, AK, FL, BL, AH, CM, LR, SS, HTS, AFT, GW, RZ, NKA, JG, DCS, and JCM collected primary data. RAF and PA performed the statistical analysis. All authors read and approved the final manuscript.