The incidence rate was 106 cases per 100,000 population. The age-specific incidence rate for Hand, foot and mouth disease was highest in the age group of 1 - 3 years old (75.04%). The major aetiologic agents were EV - 71 (52.6%), CV - A6 (21.8%). R0 ranged from 1.0077 - 2.4883. The greatest burden of Hand, foot and mouth disease is in the under five years old age range. Under present conditions, hand, foot and mouth disease will continue to persist in Haiphong. Interventions should target the high risk populations and areas.
Trang 1MODELLING OF HAND FOOT AND MOUTH DISEASE
IN HAI PHONG VIETNAM 2012 - 2016
Mbinta Fenibe James 1 , Pham Quang Thai 2 , Dao Thi Minh An 3 ,
Marc Choisy 4 , Phan Hong Hai 5 , Tran Nhu Duong 2
1 District Hospital Roua, Cameroon 2
National Institute of Hygiene and Epidemiology, Hanoi, Vietnam
3 Hanoi Medical University, Vietnam
4 UMR Univ Montpellier, CNRS 5290, IRD 224 5
Preventive Medicine Center, Hai Phong
Hand, foot and mouth disease is an infectious disease caused mainly by Entero virus 71 (EV - 71) and Coxsackie virus A16 (CV - A16) Recent outbreaks of Hand, foot and mouth disease in the West Pacific region have brought the world’s attention to hand, foot and mouth disease due to increasing morbidity and mortality The purpose of this study was to describe the epidemiological distribution of hand, foot and mouth disease cases in Haiphong 2012 - 2016 and develop and test the model of hand, foot and mouth disease in Haiphong The study was done in Haiphong using surveillance data from 2012 - 2016 All statistical analysis were done using R packages (poseid, lubridate, dplyr, magrittr, ggplot2, deSolve, bbmle, fitsir…) SIR model was used to estimate the basic reproductive number (R0) The incidence rate was 106 cases per 100,000 population The age-specific incidence rate for Hand, foot and mouth disease was highest in the age group
of 1 - 3 years old (75.04%) The major aetiologic agents were EV - 71 (52.6%), CV - A6 (21.8%) R0 ranged from 1.0077 - 2.4883 The greatest burden of Hand, foot and mouth disease is in the under five years old age range Under present conditions, hand, foot and mouth disease will continue to persist in Haiphong In-terventions should target the high risk populations and areas
Keyword: HFMD, Epidemiology, Modelling, SIR Model, Basic Reproductive number, Hai Phong
Corresponding author: Mbinta Fenibe James, District
Hospital Roua, Cameroon
Email: Mbintafenibe@yahoo.com
I INTRODUCTION
Hand, foot and mouth disease (HFMD) is a
common infectious disease that occurs most
often in children (< 5 years) but can occur in
adolescents and occasionally in adults [1] It is
caused by viruses that belong to the
enterovi-rus (EV) genus and the major aetiological
agents are the enteroviruses species A
(EV-A), mainly CA16 and EV71 [1; 2] It is
characterized by a brief febrile illness and
typi-cal skin rash, with or without mouth ulcers [1]
Aseptic meningitis, encephalomyelitis, acute flaccid paralysis, autonomic nervous system dysregulation, cardiorespiratory failure have been associated with EV71 [1; 3; 4] There is currently no specific antiviral treatment and no vaccine to protect against the viruses that cause Hand, foot and mouth disease [1] Outbreaks of enterovirus infection were reported in New York, 1972 and 1977, Austra-lia 1972 - 1973 and 1986, Sweden 1973, Ja-pan 1973 and 1978, Bulgaria 1975, Hungary
1978, France 1979, Hong Kong 1985 and Philadelphia 1987 [1; 5] Recent outbreaks of
Trang 2tion to Hand, foot and mouth disease due to
increasing morbidity and mortality [1; 5]
Before 1999, ~ 60% of encephalitis (a
com-plication of Hand, foot and mouth disease) in
southern Vietnam in children > 5 years of age,
were Japanese encephalitis [2] Since 2002,
less than 27% of encephalitis cases were
con-firmed as Japanese encephalitis, which
indi-cated that the epidemiology of viral
encephali-tis in southern Vietnam maybe changing [2] In
2005, 764 children were diagnosed with Hand,
foot and mouth disease in Ho Chi Minh City
with most cases (96.2%) being in children < 5
years old [2] In 2006 - 2007, 305 cases were
diagnosed [1] In northern Viet Nam, there was
one case in 2003 [1] From 2005 to 2007,
EV71/C5 was identified in 7 children with
acute flaccid paralysis and in 2008, 88 cases
of Hand, foot and mouth disease were
re-ported from 13 provinces [1] Since 2011, the
Ministry of Health classified Hand, foot and
mouth disease as a severe infectious disease
with outbreak potential and the disease has
been reported weekly by the national
commu-nicable disease surveillance system [4] The
first reported case in Haiphong was a 6
year-old girl diagnosed on the 17th of Apr 2011
From 2011 to 2012, the incidence of Hand,
foot and mouth disease was 524 per 100,000
population [6]
It is important to understand the dynamics
of Hand, foot and mouth disease spread
among the susceptible populations in Vietnam
This will enable policy makers to take effective
measures to curb the disease spread and
reduce the adverse impact of the disease [7;
8] One of the analytical models that can help
us understand the spread, predict the
trans-mission and control of Hand, foot and mouth
disease is mathematical modelling Mathe-matical models are used in comparing, plan-ning, implementing, evaluating and optimizing various detection, prevention, therapy and control programmes We decided to use Haiphong for modeling because it is the only city that has a complete line listing of surveil-lance data from 2012 to 2016 The objectives
of this study were as follows:
1 Describe the epidemiological distribution
of Hand, foot and mouth disease cases in Haiphong 2012 - 2016
2 Develop and test the model of Hand, foot and mouth disease in Haiphong
II METHODS
1 Study location
Haiphong has a population of 2.103.500 and surface area of 1.507.57 km² The aver-age annual population growth rate is 4.0% Haiphong city is divided into 7 urban districts: Kinh Duong, Do Son, Hai An, Hong Bang, Ngo Quyen, Le Chan and Kien An and 8 suburban districts: Thuy Nguyen, An Duong, Tien Lang, Vinh Bao, An Lao, Kien Thuy, Cat Hai and Bach Long Vi [9]
2 Data: All Hand, foot and mouth disease
cases in Hai Phong city were reported to Na-tional Institute of Hygiene and Epidemiology (NIHE), since Feb, 2011 [6] Data from Hai Phong city (2012 - 2016) was extracted from the NIHE surveillance data base The data was a line listing of all cases Subjects infor-mation collected included but were not limited
to age, residence (district), clinical status, hos-pital, results of laboratory tests, clinical grade
at hospital admission (mild or severe), date of admission [4; 6]
Trang 33 Case Definition
In Vietnam, Hand, foot and mouth disease
is defined as a brief febrile illness in children
accompanied by typical skin rash, with or
with-out mwith-outh ulcers The rash is
papulo-vesicular, occurring on the palms or soles of
the feet, or both In young children or infants
the rash may be maculopapular without
vesi-cles and may also involve the buttocks, knees
or elbows [4] A confirmed Hand, foot and
mouth disease case was defined as a patient
who had a positive RT–PCR assay for EV71
or other EV [4] In Haiphong city, monthly
throat swabs were collected from the Hand,
foot and mouth disease cases in 14/15
dis-tricts of Haiphong city by the Haiphong city
Preventive Medicine Center and sent to
En-terovirus laboratories of NIHE on ice for
etio-logical assays [6]
4 Variables
The variables used for the epidemiological
distribution of Hand, foot and mouth disease
in Hai Phong include the following: Time
(occurrence of Hand, foot and mouth disease
changes over time Cases were analysed in
days, weeks, months and years), place (this
provided information on the geographic extent
of the problem and demonstrated clusters
Districts to show area maps), Person (only age was used to describe the host characteris-tics), Virus serotype (groups within a single species of microorganisms, which share dis-tinctive surface structures)
5 Model Development
According to the formalism, we categorizes hosts within a population into three state vari-ables or compartments (Susceptible, Infected, and Recovered) [10] A state variable is a changing quantity that characterizes the state
of the system [8] A parameter is a user de-fined quantity that influences the value of the state variables [8] In this model, we defined two major parameters: effective contact rate (β), recovery rate (γ) [10]
In order to derive a simple model, the fol-lowing assumptions were made [10]: The
‘population is closed’ without demographics; All individuals are equally likely to be infected; There is homogeneous mixing, whereby intri-cacies affecting the pattern of contacts are discarded; Once infected and recovered, sub-jects were no longer susceptible to infection i.e lifelong immunity; No vaccines; Transmis-sion is frequency dependent
Given the premise that underlying epidemi-ological probabilities are constant, we get the following SIR equations [10]:
Trang 4= 0, (1) to (3) become:
The parameter γ is called the removal or
recovery rate, its reciprocal (1/γ), determines
the average infectious period S + I + R = N,
S (t) + I (t) + R (t) = 1
At equilibrium, the three equations
consti-tuting the system of equations are zeros, i.e
equations (1) to (3) will be zero as follows:
dS
dt = 0,
dt = 0, dt
From equation [5], either I = 0 or βS – γ =
0 When I = 0, there is disease free
equilib-rium, otherwise βS – γ = 0 i.e S = γ/β The
initial proportion of susceptible people must
cross this critical threshold for an epidemic to
occur This is a well-known result referred to
as the “threshold phenomenon” [10] S + I + R
= N
γ/β + I + 0 = N, Let, Therefore β/γ > 1
(R0)
6 Statistical analysis
The format for raw data was Microsoft
Excel 2010 All statistical analysis were done
using R packages provided by the R
commu-nity [11] Some of the packages used include:
poseid (ll2incidence), lubridate, dplyr, magrittr,
ggplot2, deSolve, bbmle, fitsir and EpiEstim
Rstudio includes a console,
syntax-highlighting editor that supports direct code
execution and a variety of robust tools for
plot-ting, viewing history, debugging and managing
workspace Fitsir was used to fit the model to
the surveillance data and also for sensitivity analysis
Data was collected from the routine surveil-lance system, therefore, there are several limi-tations such as reporting biases and quality and consistency of surveillance data report Case definitions may not be applied in the same way throughout the city We were as-sured that the surveillance team took some steps to prevent bias such as the following [6]: Training courses on the Hand, foot and mouth disease case definition and reporting were given to surveillance staff; Complement data gaps by gathering data from multiple sources through telephone, interviews with patients and their families or have meetings with physi-cians; Comparing sample characteristics to population characteristics were useful for ex-amining data for bias
7 Research ethics
Ethical clearance was obtained from the National Institute of Hygiene and Epidemiol-ogy
III RESULTS
1 Epidemiological distribution of cases
of HFMD
Trend of reported Hand, Foot and Mouth disease cases in Hai Phong
A total of 11.684 cases of Hand, foot and mouth disease were reported in Hai Phong during the 5 year period (2012 - 2016) This was 3.0% (11, 684/404, 338) of all reported cases in Vietnam The highest number of cases were reported in 2012 (35.6%) and lowest in 2015 (10.2%) (Table 1) Cumula-tively, the least number of cases were reported in February (3.5%) and the highest in
Trang 5April (11.5%) and September (10.2%) From
week 15 - 25 (2012) and week 80 - 92 (2013),
at least 100 cases were reported per week (Figure 1)
Figure 1 Trend of weekly incidence (2012 – 2016)
Figure 2 Weekly and monthly trends comparison (y1 to y5)
A comparison of the weekly cases between the years showed that an epidemic of Hand, foot and mouth disease occurred between the 5th and 28th week of 2012, the 10th and 40th week of
2015 and the 35th and 52nd week of 2016 Epidemics occurred at relatively different times of the year (Table 2) The peaks were April - May (2012), June - October (2013), May - August (2014), and August - December (2016)
Demographic Characteristics
The median and mean ages of reported Hand, foot and mouth disease cases were 2 years and 1.87 years respectively The age-specific incidence rate for Hand, foot and mouth disease was highest in the age group of 1 - 3 years old (75.04%) The highest number of cases in subjects less than 1 year old were reported in 2014 (9.65%) and the lowest in 2015 (4.62%) Between 1 - 3 years, the maximum number was 79.42% (2013) and minimum was 70.37% registered in 2012 (Table 1)
Trang 62012 (y1)
N (%)
2013 (y2)
N (%)
2014 (y3)
N (%)
2015 (y4)
N (%)
2016 (y5)
N (%) Cases N = 11684 4165 (35.6) 2853 (24.4) 1471 (12.6) 1190 (10.2) 2005 (17.2)
Prov-ince
Age
group
(age in
years)
Summary age characteristic Min
(0.00)
(1.00)
Median (2.00)
Mean (1.87)
3rd Qu
(2.00)
Max (30.00)
NA's (9) Incidence Rate (Cases per 100,000 persons), Av IR = 106
Virus Serotype
Major
sero-type
2012 - 2016
EV - 71 (90.1%)
CV - A6 &
A16 (3.3%
EV - 71 (38.8%)
CV - A6 (26.2%)
CV - A6 (50.0%)
CV - A16 (22.7%)
CV - A6 (63.5%)
EV - 17 (27.3%)
CV - A6 (31.2%)
EV - 71 (28.6%)
Table 1 Characteristics of reported HFMD cases in Hai Phong, 2012 – 2016
Incidence rate of HFMD in Hai Phong
Trang 7Figure 3 Spatial distribution of HFMD cases in Hai Phong province, 2012 - 2016
Do Son (316), An Duong (317), An Lao (318), Bach Long Vi (319), Cat Hai (320), Hong Bang (321), Hai An (322), Kien An (323), Kien Thuy (324), Le Chan (325), Ngo Quyen (326), Thuy Nguyen (327), Tien Lang (328), Vinh Bao (329), Duong Kinh (??)
The average incidence of HFMD was 106 cases per 100,000 population The incidence was highest in 2012 (190 cases per 100.000 population) and lowest in 2015 (54 cases per 100.000 population) (Table 1) The highest district incidence (370 cases per 100.000 populations) was recorded in Do Son district in 2012 (Figure 3)
Virus Serotype
During the period 2011 - 2016, the main enterovirus serotype in Haiphong were EV 71 (52.65%) and CV - A6 (21.8%) The main serotype in circulation during the 2011 epidemic in Haiphong was EV - A 71 (67.7%) (Figure 4 & Table 1)
Trang 80 5 10 15 20 25 30 35
%
Figure 4 Virus serotype in Hai Phong, 2011 - 2016
2 Development and testing of model of
HFMD
Considering the weekly distribution of
Hand, foot and mouth disease in 2012, there
were two epidemics The first was from 5th
week (26 cases) to the 28th week (41 cases)
and the second from the 28th week (41 cases)
to the 52nd week (28 cases) The parameters
in table 2 produced an 85.9% model fit on real
data (5th to 28th week) R0 was 1.0159
Regarding the 2nd epidemic of 2012 from the
28th to the 52nd week, a 73% SIR model fit
resulted from the parameters shown in table 2
R0 was 1.0100 In 2013, there was a single
epidemic from the 6th week (4 cases) to 47th
week (17 cases) There was a plateau from
the 13th to the 25th week and a peak on the
35th week The parameters produced a Ro of
1.0077 The model fit was 79.3% (Table 2)
Weekly distribution in 2014 revealed a single epidemic from 5th week (5cases) to 51st week (2 cases) During this period, the highest num-ber of cases were reported between the 20th and 30th weeks The model fit and R0 were 67% and 2.3381 respectively
Weekly distribution of cases in 2016 re-vealed two epidemics with almost the same peak, 85 cases and 98 cases respectively The first from week 1 (8 cases) to week 27 (14 cases) and the second from week 27 (14 cases) to week 51 (3 cases) Simulation using the parameters in table 2 produce an SIR model fit to the first epidemic of 80.6% The model fit was quantitatively imperfect By com-paring the model predictions with real data during the second epidemic of 2016, the esti-mated R0 was 1.0111
Trang 9Table 2 Summary of Parameters and basic reproductive number
2012
2016
R0 ranged from 1.0077 – 2.4883
IV DISCUSSION
The Hand, foot and mouth epidemics in Hai
Phong during 2012 - 2016 share many
simi-larities with previous large outbreaks in the
Asia-Pacific area [5] The number of cases
reduced gradually over the years to a
mini-mum in 2015 (10.2%) This decrease probably
corresponded to the introduction of
interven-tion strategies in Haiphong and diminishing
number of susceptible people [1] The number
of cases peaked between April and
Septem-ber accounting for 53.9% of all the cases A
study by Nguyen et al (2014) showed that
Hand, foot and mouth cases occurred virtually
throughout the year, with ~ 60% occurring
from May to October [4] Analysis of data from
the 2005 epidemic in Vietnam revealed a
dif-ferent trend [2] This is consistent with the
general trend that epidemics of Hand, foot and
mouth do not occur uniformly throughout the
year across Asia especially in the Western
Pacific Region (5)
The age-specific incidence rate for Hand,
foot and mouth was highest in the age group
revealed similar findings [5] Hand, foot and mouth is common in infants and children younger than 5 years old because they do not yet have immunity (protection) to the viruses that cause Hand, foot and mouth [5] The lower rate during the first year of life could be because of a relatively low frequency of con-tact with other children or to the presence of maternal antibodies [5] A study in Singapore revealed that, following the decline of maternal antibodies, the seroprevalance for EV-71 in-creased at an average rate of 12% per year in children from two to five years of age, and reached a steady state of ~ 50% in children 5 years and above [1]
The average incidence of Hand, foot and mouth was 106 cases per 100.000 population (Table 1) This is lower than the incidence rate
of 524 per 100.000 population reported during the 2011 - 2012 outbreak [6] The incidence rate decreased gradually over the 5 year pe-riod A study in Hong Kong showed that the depletion of susceptible people was the most important driving factor of the Hand, foot and
Trang 10appeared to have an impact on Hand, foot and
mouth transmission [12] The incidence of
Hand, foot and mouth varied from one district
to another, within the same district and from
one year to another (Figure 3) This indicates
that hotspots of Hand, foot and mouth were
not fixed and changed considerably across the
whole of Hai Phong from 2012 to 2016 A
study in 2011 in China by Liu et al, found a
similar pattern where there was clear variation
in incidence between districts of the city in
which they conducted their study They called
this phenomenon “spatial-temporal clusters”
and attributed it partly to differences in
cli-matic, geographic and social factors [13]
Regarding virus serotypes, the serotypes
were EV - 71 (52.65%) and CV-A6 (21.8%)
There was variation in virus serotype during
the 5 year period as new serotypes were
iden-tified Probably these serotypes were in
circu-lation previously but at very low levels or they
were introduced into Hai Phong from other
neighbouring provinces It maybe that virus
interaction with ecological, climate and human
factors lead to serotype conversion Serotype
conversion has already been documented in
some microorganisms [14] During the 2011 -
2012 Hand, foot and mouth epidemic in Hai
Phong, 55% of cases were due to EV-71 [6]
In 2005, an enterovirus was isolated from
53.8% of the cases: 42.1% isolates were
EV71 and 52.1% were CV - A16 (2) During
the course of 2008, 37.5% isolates were
en-terovirus-positive, including 27.3% with EV71
and 69.7% with CA16 [1]
Analysis revealed that the cases recorded
in Haiphong from 2012 to 2016 were not a
result of a continuous epidemic but due to
in-termittent increase above the baseline rate of
incidence During the five year period, there were six epidemics of Hand, foot and mouth disease in Hai Phong The worst and most severe occurred in the early half of 2012 (Figure 2)
Analysis revealed that the same parame-ters in the SIR model changed relatively little within the year (table 2) although they was variation between the years This implies that some level of prediction could be made about the incidence trends in subsequent years with the help of data from 2012 to 2016 in Hai Phong R0 ranged from 1.0077 (2013) to 2.4883 (2016) This variation in R0 was proba-bly due to variation in β and γ between the various epidemics Persistence of Hand, foot and mouth in this geographical area may be due to the changing virus (multiple serotypes) and host (birth, migration) ecology; global warming affecting the climate factors Failure
of control measures might also account for the persistence of Hand, foot and mouth [1; 5; 13]
In 2017, Chunqing Wu developed an SEIR mathematical model to estimate R0 Using data from the 2015 - 2016 outbreak of Hand, foot and mouth in Singapore, it was estimated that the yearly R0 was 1.1924 (2015) and 1.21462 (2016) Both of them are greater than
1, which implies that Hand, foot and mouth was prevalent in Singapore during these two years [15] Generally, estimates of the basic reproduction number range widely from 1.1 to 5.5 [5]
Our results likely overestimate the actual values because the model used was based on several theoretical assumptions that do not reflect reality (e.g the population is closed, the time scale of disease transmission, the inflow