Evidence Briefs

Provide a synthesis of the best available evidence on a variety of priority topic areas as identified by leading infectious disease experts. We systematically explore the published literature using a comprehensive search strategy to identify relevant research on infection prevention, management, and control. For more information on our search strategy of the published literature, click here.
McMaster University
NCCMT

Setting

Mar 28, 2017

Author(s): Stephanie Vendetti-Hastie, RN, CIC, Kristin Read, MPH, & Dr. Maureen Dobbins, PhD, RN
Expert Reviewer(s): Dr. Mark Loeb, MD, MSc, FRCPC & Dr. Dominik Mertz, MD, MSc

The OutbreakHelp Evidence Briefs aim to provide short summaries of the available evidence related to priority topic areas identified by leading infectious disease experts. Content for the Evidence Briefs were developed using a comprehensive and systematic search of the academic literature from inception to December 31st 2015 (more recently published information on this topic may be available here). All results were screened for relevance using pre-defined inclusion and exclusion criteria. Included articles must have met the following criteria: 1) specific to the topic of Ebola Virus Disease (EVD), 2) human research or research with real-world applicability, 3) study in a peer reviewed journal, and 4) published in either English or French. Articles identified as relevant were tagged with priority topic areas and assessed for quality using abbreviated versions of appropriate critical appraisal tools; a 5-star rating scheme was applied to articles as relevant. Relevance screening, category tagging, and critical appraisal were independently conducted by two raters and conflicts were resolved through discussion. A thematic analysis was performed on included articles by charting and then categorizing common concepts and topics discussed in the literature. Results are summarized in a narrative. The following Evidence Brief discusses human transmission focusing specifically on the acquisition of EVD by setting and drivers of transmission.

Main Message

EVD may be spread to susceptible contacts directly through a variety of exposures within the community (82%), hospital (12%) and funerals (6%); with transmission among family members constituting 81% of community transmission.

Social transmission pathways are important considerations in the management of EVD transmission.  As EVD is primarily spread through direct contact and has a long mean generation time of 15 days, control measures aimed at social distancing may be effective, including: early case ascertainment, isolation (including self), and early clinical management in hospital. Furthermore, as there are established infection control measures for organisms spread via direct contact, ongoing training and enforcement of personal protective equipment (PPE) use is recommended to protect healthcare workers and others in the hospital environment from forward transmission of EVD.

EVD acquisition by setting

A 2015 observational study of three regions in Guinea mapped the chains of transmission of EVD cases (n=193) and the most likely setting where the infection occurred (community, hospitals, funerals) using confirmed and probable case data, laboratory database information and interviews with patients and their family members/neighbours. Context specific and overall reproduction numbers were calculated. It was estimated that 82% of transmissions occurred in the community, 12% in hospital and 6% of transmissions occurred at funerals. Transmission between family members accounted for 81% (95% CI: 80-82) of transmission in the community and 86% (95% CI: 75-90) of transmission at funerals. Overall, 72% of transmissions occurred between family members. Prior to implementation of EVD control measures, the overall reproduction number for non-healthcare workers with EVD was 2.3 (95% CI: 1.6-3.2); 1.4 in the community (95% CI: 0.9-2.2); 0.4 in hospital (95% CI: 0.1-0.9); 0.5 at funerals (95% CI: 0.2-1.0). In March 2014, hospital transmission constituted 35% of all transmissions and funeral transmissions constituted 15% of all transmission. After EVD control measures were instituted in April (e.g. secured burial and opening a treatment centre), the reproduction number was reduced to less than 0.1. Hospital and funeral settings accounted for 9% and 4% of all transmissions respectively, with the remainder being transmission in the community. In the community setting the reproduction number was reduced by 50% to 0.7 (95% CI: 0.5-0.8) for patients admitted to hospital, but remained unchanged for those that were not (1.4, 95% CI: 0.9-1.9). Simulations showed that a 10% increase in hospital admissions could have reduced transmission chains by 26% (95% CI: 4-4.5) (Faye et al., 2015).

A 2015 modelling study investigated how household size and level of within-household transmission can affect epidemic risk, epidemic size and epidemic management by quarantine. When all other epidemiologic parameters were kept the same, the intrinsic growth rate, the basic reproduction number and the household reproduction number are all higher if households are larger in the early stages of an epidemic; for large households (n=6) with a moderate level of within-household transmission, the household reproduction number can be up to 65% higher than for a single person household. Case detection and isolation, followed by household quarantine were effective measures to control epidemic spread in larger households with intense within-household transmission. In a community composed of small households, however, a modest case detection probability was found to be sufficient to prevent an epidemic –even without quarantine.  Fully effective quarantine measures were shown to reduce the critical detection probability (“probability with which a case must be detected before death or recovery to prevent an epidemic” (Adams, 2016)) from 0.82 to 0.52 when household size is 6 and within-household contact rate is 5; this reduction in critical detection probability is less (from 0.34 to 0.31) in smaller households (n=2). The model demonstrated that communities composed of large households are more vulnerable to epidemics of EVD. Thus, community composition modulates the impact of control strategies (Adams, 2016).

Drivers of Transmission

Social drivers

A 2015 study examined the effect of socioeconomic status (SES) on transmission within Montserrado County in Liberia. Using case classification data for 3532 EVD cases and 1585 contacts, 324 unique communities of residence were identified, classified and analyzed to consider differences in transmission by high, middle, or low SES. Individuals from areas of middle and lower SES were less likely to seek care and had higher mortality as well as higher rates of transmission and spread of EVD cases across the region that had originated from lower SES areas. Refer to Table 5 and 6 for detailed results (Fallah, Skrip, Gertler, Yamin, & Gavlani, 2015).

 

Table 5: Outcome by SES

Factor Description

High SES n=544

Middle SES n=1044

Low SES n= 456

P-value

Number of contacts (mean ± SD)

7.41 ± 9.45

8.01 ± 8.53

10.31 ± 10.73

﹤0.001

Time to isolation (mean ± SD)

4.58 ± 5.03

4.36 ± 3.15

5.33 ± 5.97

0.247

Hospitalization [n (%)]

140 (33.98)

239 (27.86)

99 (27.89)

0.063

Mortality [n (%)]

205 (42.53)

412 (41.70)

199 (46.50)

0.240

 

Table 6: Secondary Cases by SES

Source Case

(% of the study population for each SES category)

Secondary Cases (95% CI)

Low SES

Middle SES

High SES

Risk Ratio

Low SES (17.2%)

5.45 (5.05-5.87)

1.79 (1.60-1.99)

1.26 (1.10-1.43)

2.16 (2.03,2.31)

Middle SES (35.4%)

0.65 (0.55-0.74)

1.13 (1.04-1.21)

0.91 (0.83-0.98)

0.94 (0.88, 0.99)

High SES (47.4%)

0.31 (0.26-0.36)

0.56 (0.52-0.61)

0.78 (0.73-0.82)

0.62 (0.59,0.66)

Genome sequencing was available in over 70% of EVD cases during the 2014/15 outbreak in West Africa (Scarpino et al., 2015). Using this data in combination with case sampling data and contact tracing data, investigators explored clustered transmission and under-reporting of cases in order to more accurately estimate the basic reproduction number. The authors concluded that only 58% of cases were sampled, which is considerably less than the 70% of samples with genome sequencing data available. This data suggests that the underreporting of cases of approximately 17% (maximum of 70%) is an important consideration when using modelling estimates to assess the required EVD countermeasure required for outbreak mitigation (Scarpino et al., 2015).

Reproductive number estimates

Two key parameters describing the spread of an infection are the basic and effective reproduction numbers defined as the number of secondary infections generated by an infected index. If the effective reproduction number drops below unity, the epidemic eventually stops. The reproductive number is important as it is an indication of the early growth rate for an epidemic and can be helpful in assessments of the required countermeasures such as number of hospital beds, PPE, staff and other resources. Five studies used cross-sectional modelling designs to examine reproductive numbers and to investigate transmission of EVD from survivors and non-survivors among contacts (Table 7). One study used clinical and epidemiologic data from EVD survivors and non-survivors to model disease transmission and to predict outbreak control measures that would have greatest effect of reducing overall transmission (Yamin, et al., 2015). Three cross-sectional studies used modelling to predict the reproductive numbers of the 2014/15 EVD outbreak in West Africa pre and post implementation of control measures (Lewnard, et al., 2014; Althaus, 2014; Althaus, 2015). One study examined the transmissibility of EVD using modelling and provided anecdotal information on potentially effective control measures (Nishiura & Chowell, 2015).

Table 7: Reproductive ratios, control measures and outcome

Study

Estimated Reproductive Ratio [RR(95% CI)]

Control Measures

Measure

Outcome

Yamin, et al., 2015

Survivors 1.73 (1.66-1.83)

Non-survivors 2.36 (1.72-2.80)

Average 1.73 (1.66-1.93)

Self-isolation (75% of all cases to isolate at symptom onset)

Estimated 60% of contacts would be reduced. 78% probability of disease elimination.

Lewnard, et al., 2014

Average

2.49 (2.38-2.60)

Expanding capacity of EVD treatment centers:

● 4800 beds over 4 weeks

● 2400 beds over 2 weeks

Protective kits

● 10% efficacy

● 50% efficacy

Increase case ascertainment 5-fold

Estimated number of cases averted [#(95% CI)]

● 49,559 (41,720-58,152)

● 62,220 (53,556-70,654)

 

● 4,497 (5,153-13,524)

● 30,557 (22,535-38,663)

26,746 to 75,065 (19,003-35-127) to (67,330-82,994)

Althaus, 2015

5.2 (4.0-6.7)

Implementation of control measures in hospital

Secondary transmission started to drop 28 days after onset of symptoms in index case (95% CI, 25-34)

Althaus, 2014

 

 

Guinea 1.51 (1.50-1.52)

Sierra Leone 2.53 (2.41-2.67)

Liberia 1.59 (1.57-1.60)

Implementation of control measures in each province

Reduction in transmission per day:

Guinea 0.0023 (0.0023-0.0024)

Sierra Leone (0.0097 (0.0085-0.0110)

Liberia – not determined

Nishiura & Chowell, 2015

 

 

1.7 (1.5 – 2.0)

Mean generation time 15 days

Social distancing, contact tracing, isolation

Due to the length of the generation time, the proposed IC measures may be effective in controlling outbreak spread

 

Conclusion

There is some evidence quantifying risk factors for transmission of EVD among humans. This research is based primarily on the EVD outbreak in West Africa and highlights social risk factors as well as variations in risk of transmission between different settings both before and after infection prevention and control measures are put in place. While there is evidence that EVD spreads primarily through direct contact with blood/body fluids of infected persons, particularly during later stages of illness in community, hospital and household settings in Africa, there are few studies that describe risk factors for EVD transmission among humans in developed countries. 

Modelling studies that consider variations in household size and within-household transmission as well as the potential impact of different control measures on reproductive rates may be useful to help inform infection prevention and control interventions in different contexts. An important consideration when using modelling studies is the potential for estimation bias as many modelling studies do not account for clustered transmission and the under-reporting of cases. 

REFERENCES

Adams, B. (2016). Household demographic determinants of Ebola epidemic risk. Journal of Theoretical Biology, 392, 99-106. [OutbreakHelp Star Rating: 4]

Althaus, C. L. (2015). Rapid drop in the reproduction number during the Ebola outbreak in the Democratic Republic of Congo. PeerJ, 3, e1418. DOI: 10.7717/peerj.1418. [OutbreakHelp Star Rating: 4]

Althaus, C. L. (2014). Estimating the reproduction number of Ebola virus (EBOV) during the 2014 outbreak in West Africa. PLoS Current Outbreaks, 6, DOI: 10.1371/currents.outbreaks.91afb5e0f279e7f29e7056095255b2883. [OutbreakHelp Star Rating: 4]

Fallah, M. P., Skrip, L. A., Gertler, S., Yamin, D., & Galvani, A. P. (2015). Quantifying poverty as a driver of Ebola transmission. PLOS Neglected Tropical Diseases, 9(12), e0004260.1. [OutbreakHelp Star Rating: 4.5]

Faye, O., Boëlle, P. Y., Heleze, E., Faye, O., Loucoubar, C., Magassouba, N. F., … Cauchemez, S. (2015). Chains of transmission and control of Ebola virus disease in Conakry, Guinea, in 2014: An observational study. The Lancet Infectious Diseases, 15(3), 320-326. [OutbreakHelp Star Rating: 4.5]

Lewnard, J. A., Mbah, M. L. N., Alfaro-Murillo, J. A., Altice, F. L., Bawo, L., Nyenswah, T. G., Galvani, A. P. (2014). Dynamics and control of Ebola virus transmission in Montserrado, Liberia: A mathematical modelling analysis. The Lancet Infectious Diseases, 14(12), 1189-1195. [OutbreakHelp Star Rating: 4]

Nishiura, H. & Chowell, G. (2015). Theoretical perspectives on the infectiousness of Ebola Virus Disease. Theoretical Biology and Medical Modelling, 12(1), 1. [OutbreakHelp Star Rating: 2.5]

Scarpino, S. V., Iamarino, A., Wells, C., Yamin, D., Ndeffo-Mbah, M., Wenzel, N. S., … Townsend, J.P. (2015). Epidemiological and viral genomic sequence analysis of the 2014 Ebola outbreak reveals clustered transmission. Clinical Infectious Diseases, 60(7), 1079-1082. [OutbreakHelp Star Rating: 3]

Yamin, D., Gertler, S., Ndeffo-Mbah, M.L., Skrip, L.A., Fallah, M., Nyenswah, T.G., Altice, F.L., Galvani, A.P. (2015). Effect of Ebola progression on transmission and control in Liberia. Annals of Internal Medicine, 162(1), 11-17. [OutbreakHelp Star Rating: 4.5]