Skip to main content

Table 1 Summary of key model specifications of reviewed models

From: A systematic review of transmission dynamic studies of methicillin-resistant Staphylococcus aureus in non-hospital residential facilities

Settings Nursing Homes
Articles Chamchod et al. (2012) [22] Batina et al. (2016a) [23] Batina et al. (2016b) [24]
Aims 1. Study MRSA dissemination
2. Study persistence and prevalence of MRSA
3. Study intervention controls
1. Assess MRSA epidemic potential
2. Determine conditions at which USA300 and non-USA300 could be eliminated or reduced
3. Evaluate the impact of recent antibiotics exposure on MRSA prevalence and Ro
1. Predict long-term prevalence of USA300 and non-USA300
2. Assess the influence of potential risk factors on MRSA acquisition rates and average duration of colonization
Country (model inference) Non-specific a Wisconsin, United States Wisconsin, United States
 Typeb Compartmental (deterministic);
Markov process (stochastic)
Compartmental (deterministic)
Markov process (stochastic)
Markov chain model
 Forecast period 1200/2000/4000 days 20 years to 30 years 120 months
Disease progression
 Host Residents Residents Residents
 Vector HCWs Not applicable Not applicable
 States involved among hosts Susceptible, Colonized Susceptible, Colonized Susceptible, Colonized
 States involved among vectors Decontaminated, contaminated Not applicable Not applicable
 MRSA Strains involved MRSA as a whole USA300, non-USA300 USA300, non-USA300
 Stratified by hosts’ recent antibiotics exposure No Yes Yes
Transmission pathways
  Residents to Residents Yes Yes Not applicable d
  Residents to HCWs Yes c No Not applicable d
  HCWs to Residents Yes c No Not applicable d
  HCWs to HCWs No c No Not applicable d
  Importation of colonized cases Yes Yes Not applicable d
Settings Correctional facilities
Articles Hartley et al. (2006) [27] Kajita et al. (2007) [25] Beauparlant et al. (2016) [26] g
Aims 1. Calculate the epidemiological weighte of an institution / subpopulation 1. Assess outbreak severity
2. Determine the conditions and consequences of outbreaks
3. Design interventions to control outbreaks
1. Determine effect of community dynamics on MRSA dynamics in prisons
2. Determine the effect of recidivisms on disease dynamics
Country (model inference) Non-specific f Los Angeles, United States United States
 Typeb Mathematical formula Compartmental (deterministic, stochastic) Compartmental (deterministic)
 Forecast period Not applicable 9 months 1000 days
Disease progression
 Host Inmates Inmates Community, Inmates, Recidivists
 States involved among hosts Colonized, Non-colonized Susceptible, Colonized, Infected Susceptible, Infected
 Strains involved MRSA as a whole CA-MRSA MRSA as a whole
 Stratified by hosts’ recent antibiotics exposure No No No
Transmission pathways
  Inmates to Inmates Not applicable Yes h Yes h,i
  Inmates to Staff Not applicable No No
  Staff to Inmates Not applicable No No
  Importation of colonized cases Not applicable Yes Yesj
Settings Inter-facilities
Articles Barnes et al. (2011) [28] Lesosky et al. (2011) [31] Lee et al. (2013a) [29]
Lee et al. (2013b) [30]
Aims 1. Predict long-term prevalence of facilities
2. Assess the effects of facility size, patient turnover and decolonization on MRSA prevalence
1. Determine how patient transfers affect MRSA transmission among patients in hospitals and NHs [29]:
1. Quantify how MRSA prevalence in NHs affect those in hospitals
[30]: 1. Compare different contact intervention strategies (no intervention VS only clinically apparent MRSA infections VS all MRSA carriers)
Country (model inference) Non-specific f Non-specific k California, United States
 Typeb Hybrid simulation model l Stochastic, discrete time Monte Carlo simulation model Agent-based model
 Forecast period Not explicitly stated 365 days [29]: 5 years after outbreak
[30]: 5 years after outbreak implementing contact precautions
 Facility involved Hospitals, General LTCFs Teaching hospitals (THs)m,
Non-teaching hospitals (NTHs)m, NHs
Hospitals, NHs
 Agent unit Facility Individual Individual
Disease progression
 States involved Susceptible, Persistently colonized, Colonized Susceptible, Colonized/Infected Susceptible, Colonized
 Strains involved MRSA as a whole MRSA as a whole MRSA as a whole
Transmission pathways
  Patients to patients Yes Yes Yes
  Patients to HCWs No No No
  HCWs to HCWs No No No
  HCWs to patients No No No
  Residents to residents Yes Yes Yes
  Residents to HCWs No No No
  HCWs to HCWs No No No
  HCWs to residents No No No
Inter- facility (patient sharing)
 Hospitals to Hospitals No Yes Yes
 LTCFs/NHs to LTCFs/NHs No No Yes
 Hospitals to LTCFs/NHs Yes Yes Yes
 LTCFs/NHs to Hospitals Yes Yesn Yesn
  1. Remarks
  2. a The study model was parameterized with data from the Norway, Ireland, France, Italy, and United States
  3. b The choice of continuous time versus discrete time model is not generally important for these systems, because the number of individuals is small and allows the efficient simulation of both model types. In general, equation-based compartment models (CMs) and agent-based models (ABMs) produce similar, but not exact, results [77, 78]. CMs are easier to implement than AMBs, but they rely on parsimony assumptions for objects in the same compartment; whereas ABMs can feature the heterogeneity characteristics down to an individual level
  4. c HCWs were either contaminated or decontaminated but not MRSA carriers
  5. d Pathway was not explicitly stated in this model, the probability of individual MRSA colonization state at time t had reflected the present amount of colonized in the facilities and individual current MRSA status. The current state at time t was assumed to be only dependent on their states at time t-1
  6. e Epidemiological weight indicates the level of release of newly colonized individuals into the community from the facility at an average daily rate
  7. f The study model was parameterized with data from the United States
  8. g This article was retrieved from Google search engine. The other 9 articles were retrieved from PUBMED
  9. h No classification over direct (social mixing) and indirect (sharing towels and personal items) transmission pathways
  10. i Include both inmates and recidivists
  11. j There were imported cases into the prisons from community. However, instead of presenting this importation as admission probability, the authors integrated the overall disease dynamics in the community and among recidivists, and allowed flows between individuals of the same disease states, regardless of subpopulation
  12. k The study model was parameterized with data from Canada
  13. l Each facility was treated an agent, while the disease progression within a facility was featured by a compartmental model
  14. m Lesosky divided hospitals into 2 types: teaching (bigger in size) and non-teaching (smaller in size)
  15. n It includes temporary hospital admission where beds in NH would be kept for the agent until his/her return [29, 30] or for 30 days [31]