Study area
Changsha (27°51ʹ~28°41ʹ N, 111°53ʹ~114°15ʹ E), a large city with 7.04 million people in central south China, is the capital of Hunan Province. It includes 6 districts, 2 counties, and 1 county-level city. There are a total of 4,586 hospitals, clinics, and public health departments all over the city. In this study, there were 49 secondary and tertiary hospitals included into the PSS system and 2 tertiary hospitals into the ISS system, respectively. The two hospitals in ISS system were also included in the PSS. The locations of the selected hospitals were shown in Figure 1.
The modified influenza surveillance system
The ISS, based on two sentinel hospitals (hospitals A and B), was set up in Changsha in September 2005. Hospital A, which is located in the south urban area of the city, is a tertiary hospital with more than 2,700 staffs and 1,839 hospital beds. Hospital B, which is located in the north urban area of the city, is also a tertiary hospital with more than 1,700 staffs and 1,593 hospital beds. Both hospitals have the administrative department that is in charge of the routine surveillance. Hospital A is a municipally designated hospital for the diagnosis and treatment of tuberculosis.
In 2006 and 2008, the ISS system became a branch of Hunan provincial and national influenza surveillance network respectively. In February, 2012, the HIS was adopted for the surveillance in hospital A, and was modified after June, 2013. The ISS in Changsha underwent five stages (stage 1: week 39, 2005 to week 52, 2005; stage 2: week 1, 2006 to week 52, 2007; stage 3: week 1, 2008 to week 5, 2012; stage 4: week 6, 2012 to week 24, 2013; stage 5: week 25, 2013 to week 41, 2016.). During stage 1 to 3, the two sentinel hospitals registered ILI cases manually in five outpatient departments, including outpatient and emergency departments of respiratory medicine, outpatient and emergency departments of pediatrics, and fever clinic. During stage 4 to 5, hospital B remained the manual surveillance in the same outpatient departments. ILI case was defined as “fever (axillary temperature ≥ 38°C) + cough or sore throat”[22, 23].
Differently, in hospital A, HIS was adopted into the ISS during stage 4 to 5 and was also named as “HIS (stage 1)” and “HIS (stage 2)” respectively. During stage 4, all outpatient departments of the hospital were included in the ISS, and the computer would emerge a popup window by HIS with the question that “ILI or not” if physicians diagnosed one of the 108 influenza-associated diseases based on the International Classification of Diseases 10th Revision (ICD-10). The physician should answer the question to continue the later part to treat the diseases from the patients. But we found that some ILI cases could still be missed probably because of the misunderstanding of the definition of ILI of the physician, especially if the physician was not in the department with the ISS system during stage 1 to 3. Therefore, during stage 5, the question was changed to three options: a) fever (axillary temperature ≥ 38°C), b) cough, c) sore throat. The procedure of HIS would count the ILI automatically by computing the number of “a) + b)”, “a) + c)”, and “a) + b) + c)”.
During the 5 stages, the patients who visited the outpatient departments of the two hospitals and were identified as potential ILI cases, were calculated every week. At least 5-20 throat swab samples of ILI cases per hospital per week were collected for testing the influenza virus by reverse transcription polymerase chain reaction (RT-PCR) and / or cell culture in the laboratory of Changsha CDC. The criteria for including ILI patients who were chosen to collect the samples were: a) the patients were in three days after illness onset date; b) the patients had no history of using antivirals. These sample selection and laboratory surveillance procedures were based on the National Influenza Surveillance Program (2010 edition and 2017 edition) which was announced by the National Health Commission of the People’s Republic of China. This system may monitor influenza and emerging avian influenza cases with mild symptoms or at the early stage of the infection (Figure 2). Data of the system from week 39, 2005 to week 41, 2016 were collected in our study. Because H1N1pdm was firstly emerged in 2009[24, 25], the virus was not tested in ISS stages 1 and 2 (Table 1).
The new pneumonia surveillance system
The PSS was built in Changsha in March, 2009. Pneumonia related inpatient departments in all 49 hospitals (excluding the primary health care centers and private clinics) in Changsha were enrolled into the system. This system monitors pneumonia cases among inpatient population. The public health staff in the surveillance hospitals would count the total number of monitored inpatients, pneumonia cases, severe or death pneumonia cases diagnosed by physicians and then they reported to CDC monthly. When cases were suspected as infected with avian influenza virus by clinicians, the throat swab or lower respiratory tract samples of the suspicious patients (either pneumonia cases, severe or death pneumonia cases) were collected for testing the virus by RT-PCR in the laboratory of Changsha CDC. All the surveillance procedures were performed in every month of each year. However, because H5N6 was first emerged in 2014 and H7N9 in 2013 in China[2, 12, 15, 17], the viruses were not tested from ISS stage 1 to 4, and from ISS stage 1 to 3, respectively (Table 1). The PSS system may monitor influenza and emerging avian influenza cases with severe symptoms or death (Figure 2). In this study, we collected the data of the system from March, 2009 to September, 2016.
Statistical methods
The sentinel hospitals A and B are located in the south and north in the same city. We assumed that the outpatients were from the same age group. Therefore, three indicators (d1, d2 and d3) were used to compare the difference between the two hospitals among the five stages. They were the differences of weekly number of monitored outpatients, ILI, and ILI% of the two hospitals, and were expressed as follows:
d1 = xA – xB
d2 = yA – yB
d3 = zA – zB
xA, xB, yA, yB, zA and zB refer to weekly number of monitored outpatients of hospital A, weekly number of monitored outpatients of hospital B, weekly ILI of hospital A, weekly ILI of hospital B, weekly ILI% of hospital A, and weekly ILI% of hospital B.
The Analysis of variance (ANOVA) was employed to show the d1, d2 and d3 among the five surveillance stages of the two sentinel hospitals. If there is a statistical significance, the Least Significant Difference (LSD) method will be adopted to conduct the multiple comparisons between any two stages. P < 0.05 was considered statistically significant.