Transcript Infection Transmission Dynamics in Hospital Settings

Wearable Technologies for Studying
Infection Transmission Dynamics in Hospitals
Valeriya Kettelhut, M.D., Ph.D., M.P.H.| UNMC, 2015
Estimating Potential Infection
Transmission Routes in Hospital
Wards Using Wearable Proximity
Sensors
Philippe Vanhems, Alain Barrat, Ciro Cattuto, Jean-Francois Pinton,
Nagham Khanafer, Corinne Regis, Byeul-a Kim, Brigitte Comte, Nicolas
Voirin
Data on Infection Transmissions in Hospitals


Close-range contacts are strong
determinants of potential transmissions of
infectious agents
The accurate description of contact patterns
between individuals is critical




for better understanding of the possible
transmission dynamics
for designing better infection prevention
and control measures
Problem: acquisition of reliable data on
these behaviors

Current methods of gathering data

Surveys

Diaries

Time use records
Problems with these methods

Lack of longitudinal dimension

Lack of high spatial and temporal
resolution (distance and time spent
for a contact)
Methods for Contact Data Collection

Sensor-based data collection

Wearable badge with ultra-low power radio packets: small active RFID devices

The study system was tuned to a specific distance 1.5 m when the radio
packets exchange can occur

Condition: the probability to detect this distance over a time interval of 20 sec.
should be larger than 90%

Location of the sensor: the chest Position: face-to-face

The signals are detected by the sensor and sent to a radio receiver

Definition of “contact”: two individual are in “contact” when their sensors
exchanged at least one packet during 20 sec


COMMENT: physical contact vs. being in proximity
The SocioPatterns collaboration: dataset www.sociopatterns.org
Study Setting, Design, and Data Collection

Individuals categorized by “role”

Acute geriatric unit of a university: 19 beds

Contact event: close-range interactions

Patients PAT
RN/Tech NUR

Subjects

Medical doctor MED
Admin ADM



29 Patients (Pt-HCW) 94% p.rate

46 Healthcare workers (HCW-HCW) 92%
Nurses, nutritionist, physiotherapist,
physicians, interns
5 daytime periods and 4 night periods
(Monday at 1:00 pm to Friday at 2:00 pm)
Patient data were de-identified

Measurements for each individual
(contact matrices)
1.
Number of distinct contacts per each
individual
2.
Total number of contacts for each individual
3.
Duration of each contact for each individual
Results
Table 1. Number of individuals in each class, and
average number and duration of contacts during the
study per individual in each class.
Group*
NUR
PAT
MED
ADM
Overall
Average
Number number of
of
contacts per
individua individual (SD)
ls
27
29
11
8
75
590 (470)
136 (112)
558 (341)
258 (291)
374 (390)
Average
duration
(seconds) of
contacts per
individual
(SD)
27111 (24395)
6327 (5421)
27307 (16275)
10135 (11439)
17293 (19265)
Numbers in parenthesis give the standard deviation.
*Abbreviations: NUR, paramedical staff (nurses and
nurses’ aides); PAT, Patient; MED, Medical doctor;
ADM, administrative staff.
doi:10.1371/journal.pone.0073970.t001
Table 2. Total number and duration of contacts between
pairs of individuals belonging to specific classes.
Pair*
NUR-NUR
NUR–PAT
MED-MED
NUR–ADM
MED-NUR
MED-PAT
MED-ADM
ADM-PAT
ADM-ADM
PAT-PAT
Total
Contact number Cumulative duration (sec)
5,310 (37.8%)
253,900 (39.2%)
2,951 (21.0%)
136,900 (21.1%)
2,136 (15.2%)
113,200 (17.5%)
1,334 (9.5%)
51,920 (8.0%)
1,021 (7.3%)
35,380 (5.5%)
574 (4.1%)
29,420 (4.5%)
272 (1.9%)
9,180 (1.4%)
227 (1.6%)
8,820 (1.4%)
115 (0.8%)
5,580 (0.9%)
97 (0.7%)
4,180 (0.6%)
14,037 (100%)
648,480 (100%) =180 h
Numbers in parenthesis give the percentage with respect to the total number and durations of all
detected contacts.
*Abbreviations: NUR, paramedical staff (nurses and nurses’ aides); PAT, Patient; MED, Medical
doctor; ADM, administrative staff.
doi:10.1371/journal.pone.0073970.t002
Table 3. Number and duration of contacts between
individuals in the various periods of the days,
aggregated over the observation period of 4 workdays and
4 nights
Number (%
of total)
Seconds (% of
total)
Minutes Hours
Mornings 9,060 (64.5)
426,860 (65.8)
7,114
118.6
Afternoons 4,165 (29.7)
185,790 (28.7)
3,097
51.6
Days
13,206* (94.1) 612,900 (94.5)
10,215
170.3
Nights
831 (5.9)
35,580 (5.5)
593
9.9
Total
14,037
648,480
10,808
180.1
Figure 3. Contacts matrices between classes of individuals in each
morning, afternoon and night.
In each matrix, the entry at row X and column Y gives the total
number of contacts of all individuals of class X with all individuals of
class Y during each period.
Abbreviations: NUR, paramedical staff (nurses and nurses’ aides);
PAT, Patient; MED, Medical doctor; ADM, administrative staff.
The evolution of the number of
contacts at the more detailed
resolution of one-hour time
windows is reported in Figure 2.
The number of contacts varied
strongly over the course of a day,
but the evolution was similar
from one day to another (for day
1 and day 5, contacts were
recorded after 1:00 pm and
before 2:00 pm respectively, with
very few contacts at night and a
maximum around 10–12 am.
Figure 2. Number of contacts per 1-hour periods.
SUPER-CONTACTORS: SUPER-SPREADERS
6 NUR accounted for 42.1 % of the all contacts
Conclusions and Future Work

Data can be used to explore the spread of infection through
mathematical and computational modeling


data can help to accurately inform computational models of the
propagation of infectious diseases and, as a consequence, to improve
the design and implementation of prevention or control measures
based on the frequency and duration of contacts
The possibility for HCWs to be super-contactors emphasizes the need
to reduce their exposure to infection and to limit the risk of
transmission to patients.
 HCWs could be warned against the risk brought forth by unnecessary
large numbers or long durations of contacts, especially with
patients.
An infectious disease model on
empirical networks of human contact:
bridging the gap between dynamic
network data and contact matrices
Anna Machens, Francesco Gesualdo, Caterina Rizzo, Alberto E Tozzi,
Alain Barrat and Ciro Cattuto
Study Aim

to compare different numerical simulations of the spread of an infectious
disease, where each simulation is constructed on top of a specific mathematical
representation of contact patterns, and all these representations are derived
from the same empirical data, summarized or modeled at different levels of
detail (e.g., individual-based contact network vs contact matrices)
Study Setting, Design, and Data Collection

The Department of Pediatrics


119 Individuals categorized by
“role”

37 patients (P), 20 physicians (D), 21
nurses (N), 10 ward assistants (A),
31 caregivers (C)

One week for data collection
It has 44 beds arranged in 22 rooms
with 2 beds

Children with acute diseases who do not
require intensive care or surgery

The pandemic period when several
patients with H1N1 infection were
admitted

Contact event: close-range interactions
Background

The integration of empirical data in computational frameworks designed to
model the spread of infectious diseases poses a number of challenges that are
becoming more pressing with the increasing availability of high-resolution
information on human mobility and contacts.

The integration of highly detailed data sources yields models that are less
transparent and general in their applicability.

Given a specific disease model (SEIR) , it is crucial to assess which
representations of the raw data work best to inform the model, striking a
balance between simplicity and detail

SEIR model: the susceptible, exposed, infectious, recovered model
Method

Type of data: high-resolution data on the face-to-face interactions of individuals
in a pediatric hospital ward, obtained by using wearable proximity sensors

To simulate the spread of a disease in this ped. community, an SEIR model
(with births, deaths, or introduction of individuals) was used on top of different
mathematical representations of the empirical contact patterns

All contacts between individuals and their exact timing and order were taken
into account

A hierarchy of coarse-grained representations of the contact patterns was built

The dynamics of the SEIR model were compared across these representations
Findings

A contact matrix that only contains average contact durations between role classes
fails to reproduce the size of the epidemic obtained using the high-resolution
contact data and also fails to identify the most at-risk classes.

The investigators introduced a contact matrix of probability distributions that takes
into account the heterogeneity of contact durations between (and within) classes of
individuals, and showed that this representation yields a good approximation of the
epidemic spreading properties obtained by using the high-resolution data.

The role class of the initial seed has a strong impact on the extinction probability and on the
probability of observing a large outbreak: if the seed is a ward assistant or a nurse, the
probability of a large outbreak is much larger.

In addition, assistants and nurses have an overall larger risk compared to the other role
classes. These results are consistent with literature that highlights the crucial importance of
prioritizing nurses for local infection control interventions .
Conclusions

The results mark a first step towards the definition of synopses of high-resolution
dynamic contact networks, providing a compact representation of contact patterns
that can correctly inform computational models designed to discover risk groups and
evaluate containment policies