The average rainfall by month was divided into three groups, less than 275 mm, 275–375 mm and more than 375 mm in Table 4. The highest percentage (2458; 48.8%) of all the six diseases studied were related with less than 275 mm of average rainfall. 45% (201) malaria, 43% (539) enteric fever, 48% (413) diarrhea, 52% (312) encephalitis, 52% (708) pneumonia and 53% (285) meningitis fell under this category (chi-square p = 0.010). Average rainfall of 275–375 mm and > 375mm was associated with 24% (1203) and 27% (1373) of the total disease burden respectively. Analyzing the data of 2012, in six out of twelve months, there were ≤375 mm average rainfall (Fig 5). Higher incidences of encephalitis (p = 0.001) and meningitis (p = 0.001) happened while there was low rainfall. Incidences of diarrhea (p = 0.002), malaria (p = 0.001), pneumonia (p = 0.002) and enteric fever (p = 0.002) increased with rainfall, and then gradually decreased.

Table 3 shows the association of the six diseases with the average humidity of the studied period of 2008 to 2012. We analysed the incidence of each disease in each of three groups of humidity recordings by month: less than 76.5%, 76.5–77.5% and more than 77.5% based on the mean humidity of the study site. Less than 76.5 percent humidity was associated with the highest percentage (2458; 49%; p = <0.01) of all the six studied diseases (Table 3). Average humidity of 76.5–77.5% and > 77.5% was associated with 27% (1373) and 24% (1203) disease respectively. However, in 2012, the average humidity was ≥ 76.5% in nine out of twelve months (Fig 4). Higher humidity was correlated with a higher number of cases of malaria (p = 0.0001), enteric fever (p = 0.0001) and diarrhea (p = 0.0001), but inversely correlated with meningitis (p = 0.0001), encephalitis (p = 0.0001) and pneumonia (p = 0.0001).

The two deceased children, out of seven that developed signs of severe illness after enrolment (Fig 3), were assigned influenza A and enterovirus as final diagnoses. Both received antibiotics at enrolment. The remaining five children had coronavirus and rhinovirus infection (n = 2), bocavirus, GAS infection and influenza A, respectively. There was neither any difference in early fever resolution (Table 1) among patients with (39/45; 87%) or without (504/595; 85%) a serious bacterial infection (CXR-confirmed pneumonia and/or urinary tract infection, Table 1), nor among patients treated (392/474, 83%) or not treated (143/174, 82%) with antibiotics (p>0.05 for both).

At enrolment, 500 (74%) patients were prescribed antibiotics of whom 472 (93%) received beta lactams (Table 4). Among the 152/677 (22%) patients with infections requiring antibiotics (Fig 7A and 7B) 129 (85%) received antibiotics, the most common being GAS-infection and CXR-confirmed pneumonia. The remaining 23 that were not prescribed antibiotics recovered, although one had an episode of convulsions during follow-up. Among the 525 patients without infections requiring antibiotics (Fig 7A and 7C), 294 (56%) had IMCI indication for antibiotics, a majority of them (271/294; 92%) due to IMCI-pneumonia. Conversely, 45/152 patients (30%) with infections requiring antibiotics had no IMCI indication for antibiotics. Yet 24 of these 45 (53%) received treatment.

We sampled 334 camels from nine herds (Table 1). All nine herds had at least two animals seropositive for C. burnetii (Table 2).

Based on an intraclass cluster correlation (ρ) of 0.11, there was noted a very weak cluster correlation. Univariate analysis of exposure variables revealed that herd, age group and TS were significantly associated with C. burnetii seropositivity (Table 3). Older age group and increased TS were associated with seropositivity for C. burnetii.

The final linear mixed model for C. burnetii seroprevalence was generated with ‘age group’ as a significant factor (Table 4). The serostatus of Q fever among the three age groups is as follows: 7% young (n = 56), 14% juvenile (n = 81) and 24% adult (n = 197).

Water-borne outbreaks are an acute aftermath of flood disasters, mainly as a result of contaminated drinking water supply. Intense precipitation can mobilize pathogens in the environment and transport them into the aquatic environment, increasing the microbiological agents on surface water.17-20 Chen et al.21 found extreme torrential rain (> 350mm) was a significant risk factor for enteroviruses (RR = 1.96; 95% CI 1.474–23.760) and bacillary dysentery (RR = 7.703; 95% CI 5.008–11.849). Globally, water-borne epidemics have shown an increasing trend from 1980–2006 which coincides with the increasing number of flood events.2 According to a global systematic literature review performed by Cann et al.17 the most common water-borne pathogens to be identified following flooding were vibrio spp The most common water-borne pathogens associated with heavy rainfall were campylobacter, followed by vibrio spp

Appendices A, B list published studies which have reported post-flood increases in cholera, cryptosporidiosis, non-specific diarrhea, rotavirus, and typhoid and paratyphoid.22-31 Several studies have implicated excess rainfall in water-borne disease outbreaks because of the transportation of bacteria, parasites, and viruses into water systems.22 Marcheggiani et al.18 showed a potential association between flood events and a range of water-borne infectious diseases in Italy; including, legionellosis, salmonellosis, hepatitis A, and infectious diarrhea. Reacher et al.28 performed a historical cohort study following a severe flood in 2000 in Lewes, England. The risk of gastroenteritis was significantly associated with depth of flooding in people whose households were flooded (RR = 1.7; 95% CI 0.9–3.0; p for trend by flood depth = 0.04). Additionally, an outbreak of norovirus in American tourists was linked to direct exposure to floodwater contaminated with raw sewage in Germany.29

Earlier research has shown an association between water-borne diseases and flooding in high-income countries. From 1948–1994, more than half of the water-borne disease outbreaks in the United States were preceded by heavy rainfall (p = 0.002).30 Research from Finland found that 13 water-borne disease outbreaks from 1998–1999 were associated with un-disinfected groundwater contaminated by floodwaters and surface runoff.32 Surveys in high-income countries where individuals reported their own symptoms have indicated an increase in water-borne diseases following flooding.28,30-32

Approval for the study was obtained from the Kenyan National Council of Science and Technology (NCST; permit number NCST/RRI/12/1/BS011/064) and the Institutional Animal and Care and Use Committee of the Saint Louis Zoo. Oral consent was obtained from camel owners.

There was no significant difference in coronary artery complications (Fig 3) between patients with and without infection (21 (36%) versus 37 (39%), p = 0.68). This remained true even after adjusting for IVIG resistance and number of IVIG treatment (p = 0.47). Resistance to IVIG treatment was associated with an increased risk of CA complication both as a univariate (53% versus 34%, p = 0.037), and when adjusting for the presence of infection (p = 0.039). Coronary artery dilation at time of diagnosis was found in 58 (38%) patients, similarly distributed according to the presence or absence of concurrent infection (20 (34%) versus 38 (40%) patients, respectively; p = 0.45), and persisted in 17 (11%) patients (in 5 (9%) versus 12 (13%) patients, respectively; p = 0.42). Coronary aneurysms were diagnosed in 17 (11%) patients, without significant difference between patients with and without infection neither (7 (12%) versus 10 (11%) patients, p = 0.80). While the risk of coronary aneurysm was similar between patients with viral versus bacterial infection (5 (19%) versus 2 (12%) patients, p = 0.69), patients with bacterial infection were more likely to have coronary artery dilation (13 (48%) versus 3 (18%) patients, p = 0.04).

Sampling was conducted under the permission of the owners or other responsible persons. The birds in the farms were under the supervision of veterinarians, who took different samples as part of their routine work (i.e. as screening flocks for efficacy evaluation of applied vaccines or presence of any infections) and thus part of them were used in this study. For this reason, sampling did not require the approval of the Ethics Committee.

The results of applied statistical analysis revealed the existence of significant differences between the age of birds depending on their health status (healthy, PEC and PEMS). The Kruskal-Wallis test showed that the average age of healthy, PEC and PEMS turkeys differs significantly (P = 0.036). Multiple comparison test showed that this difference was only between healthy and PEMS turkeys (P = 0.03); healthy turkeys were about 7 weeks old and PEMS about 4 weeks old. Generally the older the turkeys were, the healthier they were. Such correlation was also implied by the independence chi-square test (P = 0.007, φ = 0.31). The calculated ORs of =3.57 and 3.75 indicated that the chance of PEC and PEMS symptoms in turkeys aged 1-4 weeks are above 3.5 times higher than the chance of such disease symptoms in the older group of 5-12-week-old animals. The OR =6.92 indicates that the possibility of PEMS relative to PEC symptoms in turkeys in the fattening phase is almost sevenfold higher than in turkeys over 13 weeks of age.

In malarial patients, the body temperature was higher compared to non-parasitemic patients (39.2 °C versus 38.8 °C; p < 0.001). Consequently, malaria was positively associated with fever grade 3 (39.4 °C – 40 °C; adjusted odds ratio for both sex and age (AOR) 3.2 [1.9–5.6]) and negatively associated with fever grade 1 (<38.6 °C; AOR 0.59 [0.4–0.8]) (Supplementary Table S2). Malaria was equally distributed in all age groups. The evolution of the main clinical and biological parameters among study patients in relation to both infections malaria and H. influenza are described in Supplementary Table S3.

Laboratory results showed that anemia was not associated with H. influenzae infection, whilst the alanine aminotransferase was lower than in uninfected patients by factor 2 (18.9 IU/L versus 41.8 IU/L; p < 0.001) (Supplementary Table S3).

The main pathogen causing urinary tract infections, irrespective of sex and age, was E. coli in 52% (28/54) patients.

Figure 2 shows the frequency of the main conditions/diagnoses and their co-occurrences; with malaria and malnutrition being the second-most frequent association, seen in 12% (71/600) patients. Systemic infection (bacteria and viruses isolated from blood) occurred together with malaria in 2% (13/600) patients.

All diagnoses and their proportions are listed in Supplementary Fig. S2. In 6% (35/600) of patients, no pathogen was found.

Overall, IVIG resistance occurred in 36 (24%) patients. Children with concurrent infection had higher rates of IVIG resistance (19 (33%) versus 17 (18%) patients, p = 0.04), and higher temperature at 48 hours (Fig 1). They were also more likely to have fever > 38.5°C at 48 hours, than those without concurrent infection (16 (36%) versus 15 (20%) patients, p = 0.05). This was accompanied with higher CRP at time of diagnosis, remaining similarly higher in the first 72 hours after treatment (Fig 2).

IVIG resistance was higher in patients with proven infection on microbiologic testing and/or imaging than those without (16 (35%) versus 20 (19%) patients, p = 0.04). There was no difference in response to treatment between patients who had a proven infection on microbiologic testing and/or imaging study versus those with clinical diagnosis of infection (16 (35%) versus 3 (27%) patients, p = 0.64). Resistance to initial IVIG treatment in patients with bacterial infection was nearly double that in patients with viral infection, although not reaching statistical significance (12 (46%) versus 4 (24%) patients, p = 0.12). There was no difference in response to treatment between patients with complete and incomplete clinical criteria (25 (28%) versus 10 (17%) patients, p = 0.12).

Patients who received antibiotics during their hospital course, independent of infection status, were at higher risk of IVIG resistance than those who did not (27 (29%) versus 9 (16%) patients, p = 0.05). However, neither receiving antibiotics prior to IVIG therapy (15 (27%) versus 11 (41%) patients, p = 0.22) nor completing antibiotic course (13 (36%) versus 14 (25%) patients, p = 0.25) were associated with response to treatment.

Personal communications between key informants were conducted in conjunction with the literature review. The context of the questions included the current state of knowledge of the association between flooding and infectious diseases and potential solutions to mitigate the risks.

The risk of typhoid fever and non-typhoidal Salmonella invasive infections is highest in infants, young children, and young adults with underlying comorbidities, including severe anemia, malaria, malnutrition, and HIV infection. Moreover, recent reports from the international travelers agency showed that immunocompromised travelers, who usually follow the same itineraries of immunocompetent persons, visit countries at high risk of infections but the risk of developing travel-related diseases is five times higher if compared with that of immunocompetent persons.

However, data on groups at risk of acquiring typhoid infections are controversial and scant. Gordon showed that the immunological status cannot be associated with an increased risk or poor outcome. However, invasive diseases caused by non-typhoidal salmonellae are more frequently diagnosed in immunocompromised persons (e.g., persons with HIV/AIDS). Likewise, a study conducted in Africa did not find differences in HIV-positive patients and controls in the clinical presentation and outcomes of typhoid fever cases. In contrast, Gotuzzo and Colleagues found a rate of typhoid fever 25 times higher in HIV-positive patients than in the general population.

With the remarkable increased number of travelers from high-income countries during the last two decades, it was estimated that 1.9 million children traveled overseas every year from the United States; similarly, a significant increase of travelers (1.7 fold) was shown in Greece from 2004 to 2008. A high proportion of enteric fever cases was described in children aged 0–14 years (>26% in 2018), mainly attributed to tourism and VFR-travels. Zhou and Colleagues highlighted an increased rate of childhood enteric fever in a large tertiary care center in Canada during 1985–2013, with several cases caused by Salmonella paratyphi A and B and by bacterial strains resistant to first-line antibiotics. In Australia, 87% of the childhood cases were acquired mainly in Southeast Asia, with an annual increasing incidence from the period 2001–2005 (13 cases per year) to the period 2011–2015 (38 cases per year). Similar data were described in France, where children aged <18 years accounted for one-third of enteric fever patients, with 61% of the infections acquired in Africa.

Pre-travel counseling focused on hygiene and preventive measures could help reduce the risk of infection in individuals younger than two years, who cannot be immunized with the currently available vaccines.

Twelve percent (72/600) children met the criteria of emergency (high risk of death) based on the adapted LODS. Females were 49% (35/72), the median (IQR) age was 19 [10.8–51] months, the nutritional status was normal (weight-for-age Z-score ≥ −2.0) in 79% (57/72) patients. Among those 72 patients, only one child had homozygous sickle cell disease and another one was HIV-positive. On admission, 29% (21/72) patients had a body temperature equal or greater than 39.5 °C, and the most common associated clinical signs were vomiting and convulsions in 53% (38/72) and 42% (30/71, NA = 1), respectively. Severe anemia – hemoglobin <5 g/dL and/or hematocrit <15% – was found in 12% (7/60, NA = 12) patients. Regarding the main diagnoses; severe malaria was found in 47% (34/67, NA = 5) and lower respiratory tract infections (LRTI) in 26% (19/72) patients; the two cases of meningitis and one case of encephalitis were among them. In two patients the cause of fever remained undetermined.

Figure 3 shows the frequency of the main diagnoses, including all co-occurrences amongst them, diagnoses coded with MedDRA’s preferred terms are listed in Supplementary Fig. S3.

The two cases of death were part of this group of children presenting with danger signs.

The mean body temperature was 98.1 °F (36.7 °C) with a standard deviation of 1.1 °F (0.6 °C). The median body temperature was 98.0 °F (36.7 °C) with an interquartile range of 97.4–98.7 °F (36.3–37.1 °C). These values are consistent with previous reports (and it is worth noting that 98.6 °F [37.0 °C] is not the mean human body temperature, despite widespread belief). Overall, 2073 fevers (body temperature ≥100.4 °F, ≥38.0 °C) were observed, constituting 2.6 % of the temperature recordings (daily mean: 2.9 %, SD: 2.1 %, range: 0.0–18.6 %). A mean of 1.0 fevers was measured per thermometer each day (SD: 0.9, range: 0.0–8.0).

Hyperpyrexia (body temperature ≥106.0 °F, ≥41.1 °C) generally constitutes a medical emergency. Because hyperpyrexia is exceptionally rare, and because patients with hyperpyrexia are likely to have their temperatures measured repeatedly, we considered all measurements of hyperpyrexia occurring within the same 12 h to be of the same patient. Other than repeated measurements within the same 12 h, all measurements in the hyperpyrexic range were at least 76 h apart (mean separation: 36.3 days). This analysis revealed 25 cases of hyperpyrexia, amounting to an incidence of 1 in 2875 temperature measurements recorded in the emergency department (34.8 per 100,000; 95 % CI: 22.5–51.4). This result from an adult ED is consistent with the higher rate of 1 in 1270 patient visits for hyperpyrexia found in a pediatric ED.

One hundred-three patients under nine years of age hospitalized at the Pediatrics Department were enrolled. There were 61 males and 42 females with a male/female ratio of 1.4 to 1. The characteristics of patients are shown in Table 1. The mean age of patients was 34.9 (1–106) months, where 89.4% of patients where 5 years of age. Among the patients 18.4% were <12 months; 32% were 13–24 months of age, 39% were 25–60 months, and 10.6% were 61–106 months (Table 1). There were no statistically significant differences among age groups.

Among the patients, 42 had other significant comorbidities (e.g. chronic bronchitis, other chronic diseases and prematurity). Thirty-four patients (33%) were attending school and daycare centers.

Infectious disease control and treatment are most feasible when outbreaks can be recognized and characterized early. The emergency department (ED) increasingly acts as a gateway to the evaluation and treatment of acute illnesses, ranging from seasonal influenza to novel disease outbreaks, and has therefore become a key site for syndromic surveillance. Because the ED has the potential to both detect and warn the community about possible outbreaks or hazards, studies of the ED have multiplied as local syndromic surveillance has become technologically achievable. For example, Hiller et al. noted 24 individual systems of ED-based syndromic surveillance of influenza in a 2013 review, and additional studies have been published since [2, 3].

Yet, there are substantial obstacles to effective surveillance of disease outbreaks at the local level. Data must provide clear indications of disease while also being easy to collect and scale, attributes that are often mutually inconsistent. Influenza provides a good example of these challenges. Surveillance by the United States Centers for Disease Control and Prevention (CDC) focuses on especially clear indicators of influenza, such as virologic testing and outpatient visits for influenza-like illness (ILI) [4, 5]. Because these indicators are difficult to collect rapidly and at scale, CDC surveillance has been limited to delayed weekly reports, and concerted efforts to improve timeliness and local coverage have been discontinued. In contrast, Google Flu Trends focused on search queries, which Google can easily acquire in real-time and with extensive local coverage. However, flu search queries appear to be sensitive to factors that are misrepresentative of influenza, such as news stories, and extraneous events, such as changing search algorithms. As a result, Google Flu Trends repeatedly misestimated influenza [8–10], and the Google Flu program was recently closed down. At the same time, Google also stopped publishing a similar website that was designed to monitor dengue fever (Google Dengue Trends).

In this study, we present body temperature as a syndromic indicator that may offer a good balance of objectivity and ease of collection, and therefore might address some of the limitations to previous methods of local disease surveillance, both for influenza and for other diseases. As a clear indicator of fever, body temperature offers an objective means of surveillance for febrile diseases such as influenza, dengue, Severe Acute Respiratory Syndrome (SARS), and Ebola. Since temperature is routinely measured already, a system that allows it to be collected passively—without additional time or effort from the user—could be widely applicable. Further, body temperature is one of the few health measurements that is understood by laypersons. Consequently, it could also be applied for disease surveillance in non-clinical settings, such as schools and workplaces.

With these potential advantages in mind, we implemented a system of automated temperature collection and deployed this system in an active ED. This technical advance article presents the system’s implementation and the general features of the collected data, including comparisons with regional disease surveillance.

Streptococcus pyogenes (Group A streptococcus) is a common pathogen responsible for a number of human suppurative infections, including pharyngitis, impetigo, pyoderma, erysipelas, cellulitis, necrotizing fasciitis, toxic streptococcal syndrome, scarlet fever, septicemia, pneumonia and meningitis. It also causes non-suppurative sequelae, including acute rheumatic fever, acute glomerulonephritis and acute arthritis. Scarlet fever, characterized by a sore throat, skin rash and strawberry tongue, is most prevalent in school children aged four to seven years old. This disease was listed as a notifiable disease in Taiwan until 2007; as such, all cases of scarlet fever had to be reported to the public heath department. According to our records, however, only 9% of the medical centers, regional hospitals and district hospitals in central Taiwan reported cases of scarlet fever to the health authorities between 1996 and 1999. The number of scarlet fever cases is therefore likely to be significantly underreported. Scarlet fever outbreaks frequently occur in young children at day-care centers, kindergartens and elementary schools and also occur in adults upon exposure to contaminated food.

Genotyping bacterial isolates with various methods is frequently used to compare the genetic relatedness of bacterial strains and provides useful information for epidemiological studies. In a previous study, we used emm (gene of M protein) sequencing, vir typing and pulsed-field gel electrophoresis (PFGE) typing to analyze a collection of streptococcal isolates from scarlet fever patients and used these data to build a DNA fingerprint and emm sequence database for long-term disease surveillance. Vir typing has since been abandoned in our lab because it has lower discriminatory power than PFGE and the protocol is difficult to standardize with conventional agarose gel electrophoresis. In contrast, the PFGE protocol for S. pyogenes has been standardized in our laboratory, and a second enzyme, SgrAI, has been found to replace SmaI for analysis of strains with DNA resistant to SmaI digestion. Since PFGE is highly discriminative and emm sequencing provides unambiguous sequence information regarding emm type, we adopted these two genotyping methods to characterize streptococcal isolates and build a Streptococcus pyogenes DNA fingerprint and sequence database for the long-term study of scarlet fever and other streptococcal diseases.

The number of scarlet fever cases in central Taiwan fluctuated greatly between 2000 and 2006. Relative to the number of scarlet fever occurrences in 2000, occurrences increased in 2001 and doubled in 2002, but dramatically dropped in 2003. The number of occurrences increased again since 2004. In this study, we characterized 1,218 isolates collected between 2000–2006 by emm sequencing and PFGE. The bacterial genotyping data and the epidemiological data collected via the Notifiable Disease Reporting System (established by Taiwan Centers for Disease Control (Taiwan CDC)) were used to examine the significant fluctuation in the number of scarlet fever cases between 2000 and 2006.

The mean (standard deviation (SD)) age of all the confirmed cases was 30.5 (16.4) years. The ages did not differ significantly between patients with H1N1-2009 and the other influenza viruses (p = 0.11). Males accounted for 9 cases (56.3%), and gender was not significantly associated with H1N1-2009 (p = 0.13). A total of 13 cases (81.3%) were under medications upon arrival. Five of the 9 H1N1-2009 cases (55.6%) had taken commercially available cold/cough medications containing antipyretic substances, and one child case among the remaining four cases took an antibiotic (azithromycin) prior to arrival. These medications were started at 20 hours to 2 days before arrival. All 7 cases with the other influenza viruses were under medications: five with commercially available cold/cough medications containing antipyretic substances, one with oseltamivir and one with an antibiotic (cefcapene pivoxil hydrochloride). Medications were not significantly associated with H1N1-2009, when the antibiotics were both included and excluded (p = 0.21 and p = 0.31, respectively).

Among the 9 confirmed cases with H1N1-2009, the axillary temperature upon arrival ranged from 36.6-38.5°C with a mean (SD) of 37.2°C (0.7°C). The axillary temperature of the cases with the other influenza viruses ranged from 35.0-39.6°C with a mean (SD) of 37.3°C (1.5°C). The axillary temperature did not differ significantly between the two groups (p = 0.95; Figure 3), and the proportions of hyperthermia also did not differ significantly between the two groups for the cut-off levels of 37.5°C, 38.0°C and 38.5°C (p > 0.05 for all cut-off levels). For the cut-off levels of both 37.5°C and 38.0°C, the sensitivities of hyperthermia for detecting influenza were estimated to be 22.2% (95% CI: 0, 56.0) for H1N1-2009 and 42.9% (95% CI: 14.3, 85.7) for the other influenza viruses. Using 38.5°C as the cut-off level, the sensitivities were estimated to be 11.1% (95% CI: 0, 33.3) for H1N1-2009 and 28.6% (95% CI: 0, 57.1) for the other influenza viruses. Age and gender were not significantly associated with the proportion of hyperthermia cases among the total of 16 confirmed influenza cases using all three cut-off levels (p > 0.05 for all cut-off levels). Medications were also not associated with hyperthermia among the 16 cases, when the antibiotics were both included and excluded (p > 0.05 for all cut-off levels). Among the 9 cases with H1N1-2009, medications were not significantly associated with hyperthermia (p > 0.05 for all cut-off levels), but the proportion of hyperthermia cases was smaller among those with medications for the cut-off levels of 37.5°C and 38.0°C. For both cut-off levels, the sensitivities of fever for detecting influenza were 16.7% (95% CI: 3.0, 56.4) and 33.3% (95% CI: 6.1, 79.2) among those with and without medications (including antibiotics), respectively.

Fever is one of the most common symptoms of disease in childhood and results in psychological and economic burdens for patients and their families. The prevention of febrile diseases therefore plays an important role in child health. The most common cause of fever in childhood is respiratory infection.1 However, the evidence for putative preventive approaches for such diseases in childhood is not yet conclusive.

The custom of gargling as a preventive approach is not widespread in many Western countries. In Japan, however, health authorities have officially recommended gargling to prevent respiratory infections for more than 90 years, and almost all Japanese believe in the preventive effect of gargling.2,3 Although the effectiveness of gargling had long been unproven, a recent randomized controlled study in Japan showed that gargling with tap water inhibited the onset of upper respiratory tract infections among adults.4 Another study suggested that gargling among adults had beneficial economic effects.5 Nevertheless, the effectiveness of gargling among children remains to be clarified.

As an initial step in collecting the necessary data, we conducted an observational survey of a large number of children. Because it would have been prohibitively expensive to investigate complicated outcomes requiring diagnosis by a doctor, we focused on overall febrile disease as a proxy of respiratory infections among children. Our aim in this large-scale population survey was to examine whether gargling prevented development of fever and the incidence of sick absence among children.

Figure 2 shows the changes in mean body temperature at admission and during the 7 days after admission between Adv and Non-Adv group patients. In addition, we also compared the fever and response to antipyretic treatment between the patients in the no pathogen group and in the Adv and Non-Adv groups, the results of which are shown in Table 3. In general, Adv group patients had a much longer duration of fever after admission than the Non-Adv patients (3.2 ± 1.6 days vs. 1.9 ± 1.2 days, 2.2 ± 1.5 days, P = 0.018) and symptom onset (5.8 ± 2.2 days vs. 3.9 ± 2.5 days, 3.7 ± 2.0 days, P = 0.006). To evaluate the degree of fever, we assessed the mean temperature and number of patients to the maximal temperature at admission. The Adv group patients had higher a mean temperature at admission (37.8 ± 0.3 °C vs. 37.3 ± 0.3 °C vs. 37.3 ± 0.2 °C, P = 0.005), and more observed instances of a temperature of over 40 and 39 to 40 °C (P < 0.001). The Adv group patients took longer to attain the maximal fall in body temperature than the Non-Adv and unknown pathogen group patients at admission (10.2 ± 5.6 vs. 8.0 ± 4.5 vs. 8.6 ± 5.5, P = 0.015).

Approximately 18% of Adv group patients had no response to antipyretic treatment, which represented a higher proportion compared with that observed in the Non-Adv or unknown pathogen group patients (P < 0.001). However, the proportion of complete response to antipyretic treatment was comparatively lower in patients in the Adv group than that observed in the Non-Adv or unknown pathogen groups.

Fever is the most common chief complaint of emergency patients and is an important pathophysiological process and common symptom of many febrile diseases.[1,2] In recent years, major public health events, such as severe acute respiratory syndrome (SARS), which are mainly manifested by fever, have attracted worldwide attention.[3–5] A fever may occur in sepsis and other infectious diseases, and may also be seen in many non-infectious diseases, such as malignancy, tissue ischaemia, cerebrovascular accident and autoimmune disease.[2,6] Sometimes it is difficult to diagnose the cause of fever, such as the fever of unknown origin (FUO) and the patient's condition may deteriorate sharply. Early identification of patients at an increasing risk of death may avert adverse outcomes.

Because of the complexity of fever-related illnesses, no biomarker can definitely diagnose sepsis or predict its clinical outcome.[7] General-purpose illness severity scoring systems such as the Acute Physiology and Chronic Health Evaluation (APACHE II) often contain too many complex items or are not specific to people with fever.[8,9] With the continuous development of machine learning technology,[10,11] a machine learning approach has outperformed existing clinical decision rules as well as traditional analytic techniques for predicting in-hospital mortality of emergency department (ED) patients with sepsis.[12] This study, using big data analysis technology, aimed to explore the key factors associated with adverse prognosis of patients with febrile illness, establish an effective model to predict fatal adverse prognosis in patients with febrile disease, and provide technical support for auxiliary clinical diagnosis and treatment decision-making.

Causative agents of diarrhea were detected in 164 of 207 diarrheal calves (79.2%). The remaining diarrheal calves were placed in the following diagnostic categories: abomasal impaction (n = 6), abomasal ulcer (n = 3), and others, which included diarrhea with unknown causes and undetected causative agents owing to insufficient fecal samples.

Fourteen species of causative agents were detected in 164 of 207 diarrheal calves. Rotavirus was the most common causative agent (57/164 calves, 34.8%), followed by Eimeria spp. (52/164, 31.7%), E. coli (36/164, 22.0%), Giardia spp. (23/164, 14.0%), Clostridium difficle (C. difficile, 16/164, 9.8%), BVDV (14/164, 8.5%), coronavirus (13/164, 7.9%), Cryptosporidium spp. (12/164, 7.3%), torovirus (11/164, 6.7%), parvovirus (9/164, 5.5%), norovirus (8/164, 4.9%), kobuvirus (3/164, 1.8%), adenovirus (2/164, 1.2%), and Salmonella spp. (1/164, 0.6%).

Viruses (117/164, 71.3%) were the most common causative agent, followed by protozoans (87/164, 53.1%) and bacteria (53/164, 32.3%). Although several causative agents were detected, rotavirus, Eimeria spp., Giardia spp., E. coli, Cryptosporidium spp., C. difficile, BVDV, and coronavirus were the main causative agents in KNC.

A total of 95 from 164 calves (57.9%) were infected with a single causative agent, while 69 calves (42.1%) were infected with multiple causative agents. A total of 48 of 164 calves (29.3%) were infected with viruses alone, while 18 (10.9%) and 29 (17.7%) were infected with bacteria and protozoans alone, respectively (Table 1).

In cases of mixed infection (69 out of 164 calves, 42.1%), 53 calves (32.3%) were infected with double causative agents and 9 (5.5%), 6 (3.7%), and 1 (0.6%) were infected with triple agents, quadruple agents, and quintuple agents, respectively (Table 2). Eimeria spp. (32/164, 19.5%), rotavirus (29/164, 17.7%), and E. coli (23/164, 14.0%) were the most commonly detected agents in calves with mixed infections (Table 3). Most causative agents were related to mixed infection rather than single infection.

We compared the laboratory and radiologic findings between the Adv and Non-Adv groups (Table 2). The percentage of patients having leukopenia and thrombocytopenia was higher in the Adv patients P < 0.001), while leukocytosis was more common in the Non-Adv group patients (P = 0.035). The levels of infection markers, such as C-reactive protein (CRP) and procalcitonin showed no difference between the two groups. In addition, total bilirubin and creatinine levels showed no significant difference between the two groups.

Possible causative agents were identified in 100% of the Adv group patients and in 72.8% (134/184) of the Non-Adv group patients. In some instances, Adv group patients had coinfections with viruses, such as rhinovirus (n = 5), influenza A virus (n = 4), respiratory syncytial virus (n = 1), and parainfluenza virus (n = 1). Bacteria or combined etiologies were more common in the Non-Adv group patients. Rhinovirus (40/184, 21.7%) was most commonly identified as the pathogen in the Non-Adv group patients. The most common bacterial pathogens were Streptococcus pneumoniae in the Adv group patients (11/67, 16.4%) and Haemophilus influenzae in the Non-Adv group patients (52/184, 28.3%).

The most common radiological feature was ground-glass opacity with consolidation in the Adv group and consolidation in the Non-Adv group (P < 0.001). Unilateral distribution was dominant in both groups (83.5% vs. 72.7%), however multilobar (≥ 3 lobes) involvement was more common in the Non-Adv group (9.0 vs. 22.3%, P = 0.015). The presence of pleural effusion was not significantly different between the two groups.