**4. Discussion**

This study provides a long and continuous evaluation of the prevalence of schistosomiasis in China. Using disease data from 1981 to 2017, we found the changes of infection rates of humans within each historical phase were on account of different and complex causes, particularly social, economic, political, and health aspects. The schistosomiasis control strategies were suitable for the respective situations back then. Compared to most other studies concentrating on a single epidemic period or data from specific time points, infection rates and control strategies of schistosomiasis in this study are

evaluated in a more comprehensive, systematic, and direct way. We confirmed that snail control and chemotherapy in combination has better effect than chemotherapy alone. In addition, we esteemed that the morbidity control measures taken in each historical period were appropriate and suited to the changing disease and social-economic status situations in China. We believed that WBLP is a successful example of disease control aided by foreign funds. Besides, we concluded the integrated strategy has greatly reduced the disease burden of schistosomiasis in China.

After different control measures were stopped in 1984, the prevalence in Sanlian village started to rise the next year, but still remained lower than that of the Guifan village until 1988. This suggests chemotherapy combined with snail control achieved robust and longer-term effects in reducing the prevalence. This conclusion corroborates control projects undertaken in other countries [15,16]. However, from 1987 to 1992, the prevalence oscillated in both villages, which was similar to other regions in China at that time. There are several possible reasons. First of all, since 1978, China's rural areas have begun to shift from a centralized system to a contracted responsibility system for joint production. The mass extermination of snails is no longer practical. Moreover, as a result of responsibility privatization and the subsequent rise of labor cost, the previous schistosomiasis control measures become too expensive. Second of all, with the reform of China's social and economic system in the 1980s, the source of schistosomiasis control budgets shifted from the previous three-level system (county, provincial, and the central governments) to the local governmen<sup>t</sup> only, significantly intensifying the burden on local finances and therefore limiting corresponding expenditure on schistosomiasis control. Third of all, previous drug eradication of snails and environmental modification brought many problems such as environmental pollution and flooding, further increased economic burden and social impact [17]. Last but not least, praziquantel, a very effective chemotherapy drug started to be introduced to and widely used in China in the late 1980s [3], the Chinese governmen<sup>t</sup> and the public dedicated less attention to schistosomiasis, which also partly led to the oscillation in prevalence.

In 1992, large-scale drug treatment, health education and snails control were implemented in both villages, which led to obviously prevalence declined in the next two years. In 1990s, following the WHO's recommendation that developing countries should utilize chemotherapy to achieve better morbidity control [18,19], a wide application of praziquantel, indirect hemagglutination, and the Kato-Katz method were initiated. This marked the starting point when China's schistosomiasis control strategy switched from environmental modification to praziquantel-based control. Particularly, in 1992, with the help of the WHO, China began to implement the World Bank Loan Project (No. P003624) for "Infectious and Endemic Disease Control Project" [20,21], which provided 71 million dollars to help and support the government's financial expenditure on schistosomiasis control, the application of parasite diagnostic techniques and the promotion of chemotherapeutic drugs [22]. Following a sharp decline in 1992–1994, the prevalence in the two villages rebounded in 1995 due to flooding in Guichi district. On the contrary, in 1997, the drought in Guichi decreased the risk of infection and therefore the prevalence. Similarly, prevalence peaked again in 1998 and 1999 as a result of severe flooding in China's Yangtze River Basin that year [23]. Many new snail areas also appeared after the flood and this aggravated the disease burden [24]. As shown in Figure 2, prevalence in two villages oscillated at a high level from 1995 to 2004 even with continuous praziquantel-based control, indicating human chemotherapy alone might not be sufficient to completely eliminate schistosomiasis. This was also supported by another study [25]. Fluctuation of prevalence in the late WBLP period was also confirmed by another study [26]. There are four possible reasons for the changing prevalence at that time. Firstly, the reform of governmen<sup>t</sup> institutions and the loss of well-trained technicians caused difficulties in normal operation of schistosomiasis control and budget reduction. Secondly, with the termination of the WBLP, the public's compliance [27] with drug treatment had also decreased. Thirdly, although the prevalence and incidence of acute schistosomiasis dropped to a relatively low level in the early stage of this period [28], praziquantel could not stop transmission in high-epidemic areas and reinfection [3]. Lastly, natural factors such as floods also contributed heavily to the prevalence fluctuation.

From 2005 to 2017, the infection rates in both villages decreased significantly. This might result from four aspects. First of all, in 2004, Chinese governmen<sup>t</sup> set up four priority control programs of communicable diseases including schistosomiasis, AIDS, tuberculosis and hepatitis B [14], and schistosomiasis received an increased attention. Second of all, Chinese governmen<sup>t</sup> launched the "National long-term plan for control of schistosomiasis (2004–2015)" program [29] in response to the rebounded trend of prevalence, which marked the beginning of the integrated control strategy in the country. Third of all, as the governmen<sup>t</sup> increased its public health budget for schistosomiasis, the governmen<sup>t</sup> transfer paymen<sup>t</sup> program was started and the national health system was strengthened. Lastly, the comprehensive control strategy, including praziquantel treatment of human and animals, molluscide application, health education, sanitation improvement, removal of bovines, etc., were implemented in our study area in 2006 [30] which greatly helped to reduce the disease burden of schistosomiasis. This result is consistent with the conclusions of many other studies [6,30–38]. However, several issues remain as well. Firstly, the comprehensive control strategy is not able to further reduce the infection rate to less than 3% in some areas [30,39]. Secondly, although removal of agriculture main livestock, such as bovines, sheep, and horses, can reduce human prevalence to a large extent [24], the existing control measures need to be adjusted in China since 45 species of mammal can be infected by *S*. *japonicum* in total, including but not limited to dogs, cats, and rodents [40–42]. Thirdly, the broadly used Kato-Katz method has relatively low sensitivity, particularly for subjects with low infectiosity [43,44]. This could lead to false negative results. Recently, several studies confirmed that Helmintex method is a more sensitive egg detection procedure than Kato-Katz method and other diagnostic tests like the point-of-care immunodiagnostic for detecting schistosome cathodic circulating antigen (POC-CCA) and Saline gradient [45,46]. Perhaps widely used of Helmintex in China is feasible and necessary. Fourthly, impact of water conservancy project on schistosomiasis control still remains controversial, due to many influencing factors such as water level, temperature, and plants change [47]. Lastly, environmental change, migration of population, and effectiveness of drug treatment could also influence the transmission of *S. japonicum*. Further investigations and effective long-term policies in controlling these factors are needed.

The main limitation of our study is that the two villages are both typical bottomland areas and cannot represent all types of epidemic areas. In addition, a comparative study of the two villages was only run from 1981 to 1984. Therefore, there is no comparison group afterwards. Moreover, the study used three different diagnostic methods: two slides Kato-Katz, indirect hemagglutination assay (IHA), and three slides Kato-Katz, however, infection rate tested by IHA cannot be adjusted with its sensitivity and specificity [10].
