Situations of work-related diseases and injuries among agriculturists in the upper northeast regions of Thailand

Background: Agriculturists exposed to health hazards are affected by increased occupational disease. This retrospective study aimed to investigate situations of work-related diseases and injuries among agriculturists in upper northeast Thailand. Methods: The secondary data of international classification of diseases 10 th revision (ICD-10) case reports of occupational disease among farmers, from the database of the Health Data Center (HDC), were used. The registered farmers data was collected as a dataset from the provincial agricultural office and the data of ICD-10 code utilised from the hospital information system (HIS) of healthcare services in Udon Thani and Roi-Et provinces, which was extracted for cases of work-related diseases and injuries of registered agriculturists. The annual morbidity rate of occupational diseases was analysed and presented at a rate per 100,000 farmers. Results: Among farmers in the HDC database, lung disease, which was not reported as occupational disease of the HDC database, was the highest ranking of all diseases, followed by work-related musculoskeletal disorders (WMSDs), noise- and heat-related diseases, and pesticide toxicity, respectively, while the injury rate was as high as that of WMSDs. Those morbidity rates of Roi-Et and Udon Thani provinces were representative of the morbidity ranking of diseases of the nation and had increasing trends from 2014 to 2016. The number of farmers in the HDC database did not consistently reflect the number of registered farmers in the agricultural database. Conclusions: Situations of work-related diseases and injuries discovered among registered farmers reflect the health problems of Thai agriculturists, and the underestimation in the reported disease rate in the health database is explained by big data analysis, which showed that work-related cases with an identifying code of Y96 had rarely been reported among agriculturists. Therefore, Thai agriculturists should be supported in registration with occupational diseases and injuries surveillance as holistic healthcare


Introduction
Agriculture is an essential source of livelihood for developing countries. However, it can lead to poor health and is linked to the main causes of health hazards of agriculturists affected by infection, injuries, chronic exposure, and increased occupational disease. 1 According to a global report on occupational diseases, the highest prevalence of such disease was work-related musculoskeletal disorders (WMSDs), which ranged between 55.6-97.2%. [2][3][4] The highest reported prevalence was in Korea (97.2%), 2 followed by Saskatchewan province of Canada (85.6%), 3 and Ireland (55.6%). 4 The next most prevalent disease was heat-related disease, which ranged from 35. 6-72.3%. 5-8 The highest reported prevalence was found in Eastern North Carolina, USA, 5 followed by Northeast Italy, 6 and North Carolina, USA. 7,8 Additionally, prevalence of hearing loss from noise was reported at levels of 16.9-36.1%, 9,10 which were 16.9% 9 in Korea and 36.1% 10 in the USA, while work-related injuries were reported at a prevalence of 69.0% in Nepal. 11 Infectious disease was reported in ranges between 0.7-13.5%, [12][13][14] where brucellosis cases in Greek farmers and livestock breeders corresponded to an annual incidence rate of 7.1 per 100,000 population, 15 which were 8.9% in the Western Cape of South Africa. 16 The pesticide toxicity was reported as the rate of 8.8-17.0%, which were 8.8% in China, 17 and 17.0% in the Sunsari District of Nepal. 18 Chronic lower respiratory tract disease was also reported as chronic obstructive pulmonary disease (COPD) and was found in a systematic review to be from 3.0% to 68.0%, 19 as in France, where the rate of COPD was 9.5%, 20 and South Africa, where the rate was 3.0% for asthma. 21 From the survey results in 2019, it was found that the total number of employed persons in Thailand was 37.5 million persons. The number of workforce members who had no social security or were informal workers was about 20.4 million, or 54.3%. By region, those in informal employment working in the Northeast made up the largest proportion (34.9%), followed by Central Thailand (23.4%), the North (20.9%), and the South (14.0%). 22 When the economic activities of those in informal employment were considered, it was found that more than half of all informally employed persons worked in the agriculture sector (about 11.5 million, or 56.4%). 22 In Thailand, the Health Data Center (HDC) (https://hdcservice.moph.go.th/hdc/main/index.php), which is the central health database system of Thailand, collects and analyses the disease surveillance system data of all healthcare services under the provincial public health office by using 43 files, or standard structured health data classified with ICD-10 codes for diseases and disorders, and records the patient's information every time they receive services from officials. The health database of the HDC showed that the highest prevalence of disease between 2014 and 2016 was for work-related injuries (about 0.6%), followed by WMSDs (about 0.3%), noise-induced hearing loss and heat-related disease (0.2%), and pesticide toxicity (0.04%). There was no statistical report on work-related diseases which included respiratory symptoms or lung disease, and infectious disease on the health database of the HDC. In 2019, the prevalence of chronic lower respiratory tract disease was reported of 15-59 years old population to be 1.8%. 23 The previous retrospective study in 2016, which showed a higher morbidity rate of all cases in the research from the Nongbualamphu province of Thailand, found the highest prevalence was for WMSDs (21.7%), followed by heat-and pressure-related diseases (5.2%). 24 The pesticide toxicity was reported in the range of 0.14-0.18% of registered farmers, 24,25 which can be compared to the previous prevalence rate of 48.48 per 100,000 registered farmers in Southern Roi-Et. 26 The study also indicated a prevalence rate of infectious disease at 0.51%, 24 with chronic lower respiratory tract REVISED Amendments from Version 1 This updated version was improved from the first or previously published version without changing of the results and the main methodology. The major difference from this version and the first version were that of rearrangement of paragraphs in the introduction section. The background of occupational diseases and injuries and the reports of their prevalence from the previous studies from the global literatures after moving up to beginning of this section. More given explanation of the role of HDC (health data center) in the health surveillance system of Thailand, and the methods of analysis and prevalence rate shown in citations from our preliminary studies in Thailand after grouping this background to move to the second or the latter section of the introduction. Another major change was in the discussion, according to the conclusion of results pointed to discussion that focused to less emphasis of interprovincial comparison, but representative the rate and trend of health problems of Thai agriculturists, which was a conclusion in an abstract and the conclusion. The strong point of the paper is the linking the identified cases to the full farmers population, registered agriculturists, and the analysis outcome of morbidity rate was showing the magnitude of risks of the agricultural farmers, which is a large part of the workforce in not only in Thailand but throughout globally. Even in the developed countries hazards in agriculture are usually underestimated. We have very much appreciated for all suggestions from the reviewers, which could make an improvement of this paper on those two sections to convey important message of occupational diseases and injuries surveillance system in Thailand and globally.
Any further responses from the reviewers can be found at the end of the article disease (1.11%) and skin disease (1.5%) 24 as additional diseases which occurred. The case-control study of registered agriculturists in Khon Kaen confirmed that the noise-induced hearing loss cases from the 43 health files data had been diagnosed through the audiometer test of those agriculturists. 27 Work-related injuries had a prevalence of 20.1% among the farmers of the Phayao province 28 and the three-year morbidity rate of heat-related illness in our most recent report (2020) was 13.5 per 100,000 farmers, or 0.014%, in registered agriculturists of Khon Kaen province. 29 Those work-related diseases and injuries in Thailand were identified by the International Statistical Classification of Diseases and Related Health Problems 10 th revision (ICD-10) (https://icd.who.int/browse10/2019/en) 30 and the specified Y96 external code (work related condition). Nowadays, however, some cases of agriculturists are not practically identified, and this leads to inconsistency in data as above. Hence, healthcare for agricultural workers in Thailand is not covered by labour legislation and as a result they only access services from the National Health Security Office (NHSO) (https://eng.nhso.go.th/view/1/Home/EN-US), 31 without claiming compensation for diseases related to agricultural activities. Although occupational disease case reports are important for enabling the surveillance system to carry out national health policy among agriculturists, there is still no representative morbidity rate of the upper northeast of Thailand. Therefore, this study aimed to investigate situations of work-related diseases and injuries among agriculturists in the upper northeast of Thailand by using the case study of the Udon Thani and Roi-Et provinces.

Ethical approval
This study was approved by the Human Research Ethics Committee of Khon Kaen University (No. HE592154). The underlying data for this study is secondary data collected by the Health Data Centre and agricultural registries and is restricted due to ethical and data protection considerations. Under the ethical approval from the Human Research Ethics Committee of Khon Kaen University (No. HE592154), participants' data was not allowed to be shared and the secondary data of health dataset and registered farmer dataset were utilized under restriction. The individual information or further reused inclusion data is not permitted under the ethically approved conditions for human research. This was confirmed under the data protection of human/participant privacy as no individuals gave explicit written consent that their identifiable data can be made publicly available.

Study design
This study was a retrospective analytic study that used secondary data from three sources, which were 1) The health database of the Health Data Center (HDC) of Thailand, between 2014 and 2016, for the number of cases of occupational disease among agriculturists (by occupational code) who visited public healthcare service providers as details in HDC data collection and public health region seven (PH region 7) and public health region eight (PH region 8), and Roi-Et and Udon Thani provinces in the upper northeast regions of Thailand; 2) The secondary dataset of the list of cultivating farmers who had registered with the provincial agriculture offices of Roi-Et (http://www.roiet.doae.go.th/) and Udon Thani provinces (http://www.udonthani.doae.go.th/) between 2014 and 2016 as in the report Form 04-2; and 3) The secondary data of ICD-10 utilised from the hospital information system (HIS) database as detailed in the HIS data collection that used for routine care services in primary care units, secondary, and tertiary hospitals was selected from the provinces of the pilot study areas (Roi-Et and Udon Thani).

Population and sample size
The sample size of the retrospective descriptive study was the total number of registered agriculturists of the studied area and all cases of occupational diseases who were farmers visiting healthcare units for health services at the studied areas of the upper Northeast of Thailand, during the three years from 1 st January, 2014 to 31 st December, 2016. The inclusion criteria for farmers from the HDC were those registered under the occupational codes shown in Table 1. The registered agriculturists from a provincial agriculture office of Roi-Et and Udon Thani provinces who were included as the population in this study met the inclusion criteria of cultivated farming classifications according to agricultural activity during 2014-2016.
For analysis of the morbidity rate of occupational diseases among agriculturists in the upper northeast region, this study used two provinces as representative of the upper northeast of Thailand. The agriculturists from Roi-Et province were representative of PH region 7 and those of Udon Thani province were representative of PH region 8 of Thailand; they were chosen to represent the healthcare service visits for all cases among farmers in provinces of the northeast of Thailand and be compared to those of Thailand as a whole. All registered agriculturists of the two provinces were the representative population of agriculturists in those provinces in Thailand as the nature of the impact from agricultural work was found to be similar across from other areas of Thailand. 32

Data collection HDC data collection
The health database reports for national and PH region 7 and PH region 8 levels in the northeast region of Thailand were collected by using the ICD-10 code of occupational disease used in reports to the HDC 23,33-37 from 2014 to 2016. The data collected were the total number of planting farmers according to occupation code (Table 1) who had accessed healthcare in the area of interest and the number of cases of occupational disease among farmers who met the inclusion criteria for planting farmers reported to the HDC, with regard to four occupational diseases and injuries classified by icd-10 23,24,33-37 ( Table 2). The total number of farmers registered nationwide with the Ministry of Agriculture and Cooperatives and provincial agriculture offices was collected from the agricultural database. 32 Provincial agriculture offices of Roi-Et and Udon Thani provinces data collection Those data of the ICD-10 code and the registered agriculturists were also collected as a big dataset for Roi-Et and Udon Thani provinces as representative provinces in the upper northeast regions of Thailand. The secondary data that was collected from a provincial agriculture office was a big dataset with the number and lists of registered agriculturists in the pilot study areas (Roi-Et and Udon provinces) during 2014-2016. The record was explored and copied into an excel file by the information technology personnel who were responsible for the database of a provincial agriculture office with permission for the use of the data in this study. Reports included the identification number, name, gender, house address, HIS database or 43 health files data collection Another big dataset collected from the provincial public health offices was the ICD-10 code record of the health standard data structure (43 health files of file 10, file 15, and file 19) on cases visiting Primary Healthcare Units (PCUs) and the secondary or tertiary hospitals in Roi-Et and Udon provinces for healthcare services in 2014 to 2016). Health-related information of registered agriculturists regarding cases with ICD-10 codes, was listed in the 43 health files of the health standards data structure and standards used for the hospital information systems (HIS) of the hospitals in Roi-Et and Udon Thani provinces: file 10diagnosis (opd), file 15diagnosis (ipd), and file 19surveillance. The dataset of ICD-10 codes related to occupational disease and injuries data in an excel file or an SQL file were explored and copied from the provincial health database by using the permission code available only to the information technology personnel of the organisation. Data of ICD-10 code with access permission and restricted use was extracted from the 43 health files data based on identifiers of individuals registered as agriculturists with the Provincial Agriculture Office of Roi-Et or Udon Thani and sorted by using the hardware and installed software of the computer, which were a CPU @ 1.99 GHz, installed RAM of 16.0 GB, system type of a 64-bit operating system, x64-based processor, and software including Windows 10 or equal to a higher version and Microsoft Office (Excel program) 2016 licensed by Khon Kaen University, or equal to a higher version. The data was screened for duplicate registration of farmers or disease accounts, and they were removed for duplication; it was shown that the final number of registered agriculturists in the period of 2014 to 2016 in Udon Thani province was 154,478, while that in Roi-Et was 207,465 farmers. The ICD-10 codes met the inclusion and exclusion criteria outlined as shown in Table 2 for occupational disease and injuries were considered for analysis of occupational diseases among the registered agriculturists. The collaborated data of registered farmer and accounted ICD-10 codes among the registered was explored and used for the morbidity rate analysis per 100,000 registered farmers.
Data analysis STATA version 10.0 software (https://www.stata.com/) (StataCorp LLC: College Station, TX) was used for all statistical analyses in this study. A freely accessible alternative software which can be used to complete the statistical analyses in this study would be program R for statistical computing (https://www.r-project.org/). Categorical data were presented as numbers and percentages. To estimate the annual morbidity rate of an occupational disease with a confidence interval (95% confidence interval) among registered agriculturists from 2014-2016. The data of occupational diseases among planting farmers or registered agriculturists was analysed for morbidity rate as per the following formula: 24 Morbidity rate ¼ Number of cases reported among farmers or registered agriculturists in an area of interest Total number of farmers or registered agriculturists in an area of interest Â 100, 000 In case of missing codes of the 5 th digit (activities, 2) for work-related injuries or Y96 for work-related musculoskeletal disorders (WMSDs), they were also countable for the ICD-10 cases of occupational diseases or injuries among registered farmers from provincial agriculture office by considering the main ICD-10 code.
The underlying data ( For the number of farmers from HDC database, the number of HDC farmers varied upon types of diseases or injuries as well as the number of cases of diseases as shown in Table 3. The highest number of farmers accessed healthcare due to cases of noise and heat-related disease in 458) and Udon Thani (1,324,168). According to the HDC, the number of farmers who accessed healthcare services for injuries in each year varied; it was shown that in 2016, there was a total of 12,580,864, 663,729, and 41,350 farmers who visited healthcare services nationally, in Udon Thani province, and in Roi-Et province, respectively. The total number of registered farmers in Thailand recorded by the Ministry of Agriculture and Cooperatives in 2016 was 4,178,064. 32 The morbidity rate of occupational disease and injuries in Thailand, i.e., WMSDs, noise-and heat-related diseases, and pesticide toxicity and injuries (per 100,000 farmers), increased between 2014 and 2016, according to the HDC database. According to national results, the results of PH region 8 and Udon Thani supported this trend of increasing morbidity rate (as shown in Figure 1). Noise-and heat-related diseases and injuries have been increasing every year in all regions, which is also supported by the rates from Roi-Et province and PH region 7.
Morbidity rate of work-related disease and injuries among farmers who visited healthcare service providers This study analyzed the rate of chronic lower respiratory symptoms (J40-J47) as one category of lung diseases among farmers between 2014 and 2016. This was analysed in this manner following the categories of the top five prevalence of occupational diseases among crop planting farmers 24 as in Table 2. The disease was not directly reported as representative of a kind of occupational disease classified in the HDC database, but there was a reported data available as categories by age groups for 5-9 years, 10-14 years, 15-59 years, and 60 years or more, which could be accountable for the cases of workforce ages as agriculturists. 22 In 2016, the national rate of lung disease was 1,129.0 per 100,000 farmers, and the highest rate was found in Roi-Et (1,527.2), which was higher than that of the national total. The rate of lung disease was followed in order by the rates of injuries, WMSDs, noise-and heat-related diseases, and pesticide toxicity, respectively, and there was a higher morbidity rate in Udon Thani compared to Roi-Et. In Roi-Et, injuries were found to be at the highest rate (319.4 per 100,000 farmers) in 2014, which was followed by WMSDs (41.0), pesticide toxicity (11.4), and noise-and heat-related diseases (6.8), while the top-ranking rate of Udon Thani was that of WMSDs (405.3), which was followed by those of noise-and heat-related diseases (47.4), and pesticide toxicity (12.8), while the injuries rate was 400.0. In 2015, apart from injuries, lung disease (1,359.3 per 100,000 farmers) had the top-ranking morbidity rate of diseases among farmers in Roi-Et, followed by those of noise-and heat-related diseases (67.3), WMSDs (62.7), and pesticide toxicity (14.5), according to the HDC. Meanwhile, in Udon Thani, the morbidity rate of lung disease was the highest rate (831.6), followed by WMSDs (296.8), noise-and heat-related diseases (88.0). A similar trend was found during 2014-2016, in that the morbidity rates of Udon Thani, or PH region 8, in this year were mostly close to the national rate. In 2016, the morbidity rate of chronic lower respiratory symptoms (lung diseases) was the highest among all diseases and injuries in Roi-Et (1,527.2 per 100,000 farmers) and was also shown to be the highest among all diseases and injuries when compared to other regions and the nation ( Figure 1, see underlying data:  Even though the number of cases of diseases was stable, the number of people who had a recorded occupation of farmer in the HDC database and the number of registered farmers were different. Udon Thani as well as the nation always had a higher recorded number of farmers in comparison to the registered number of farmers, which contrasted with Roi-Et, as shown in Table 3.
In 2016, all rates of Udon Thani were higher than those of Roi-Et, i.e., with a 5.89, 1.16, 1.89, and 1.20-times higher rate of WMSDs, noise-and heat-related diseases, pesticide toxicity, and injuries, respectively.
Although the number of registered farmers of Roi-Et Province was higher than that of Udon Thani, the recorded cases or number of farmers to health service visits were lower compared to Udon Thani, for all diseases and injuries ( Table 3). The morbidity rate of disease was higher in Udon Thani in comparison to Roi-Et. It was found that the morbidity rates of the nation and Udon Thani Province had been increasing similarly.
When comparing the rate of occupational disease and injuries among registered farmers in 2016 to the rate of farmers who visited health service providers, the higher rate was found in the registered farmers for WMSDs, noise-and heat-related diseases, and injuries (pesticide toxicity was scarcely different), especially in Roi-Et, where the rate was 27.58 times higher for injuries and 16.32 times higher for WMSDs. Regarding those who identified as farmers when visiting Udon Thani health service providers, they had a 1.76 times lower rate of WMSDs and a 15.83 times lower rate of injuries than registered farmers. Similarly, the rates of noise-and heat-related diseases among those who identified as farmers in Udon Thani and Roi-Et provinces were about two times lower in the HDC database, in comparison to the rates of registered farmers.

Morbidity rate of work-related disease and injuries of registered farmers visited health care providers in 2016
According to the secondary dataset of the health standard data structure (43 health files) and the number of agriculturists registered to the provincial agricultural office during 2014-2016, without duplicated registration of farmers or disease    Regarding the characteristics of those cases of registered farmers in Roi-Et and Udon Thani provinces, more than 60% were female. The largest proportion of cases in Roi-Et were working on rice plantations (98.26%), followed by cassava (1.12%), sugar cane (0.19%) and rubber (0.37%) plantations. Regarding cases in Udon Thani, most cases were working on rice plantations (89.56%), and the next highest proportions were working on the following plantations, i.e., cassava (4.97%), sugar cane (4.56), rubber (0.24%), and soybean/corn (0.67%). Almost 80% of them worked in relatively small farming areas of less than 4 acres. Most of them (about 70%) produced less than 10 tons (see underlying data; The results of morbidity rates in 2016 are presented in Table 4. The ranking of morbidity rates of work-related disease among registered farmers was as follows: lung disease, WMSDs, noise-and heat-related diseases, and pesticide toxicity, i.e., 2781. 7,754.3,221.7,278.6,479.1,207.7,and 26.5 in Udon Thani, respectively. The rate of work-related injuries was more than four times higher compared to all rates of other diseases in Udon Thani and Roi-Et, as shown in Table 4.
The trend of work-related disease and injuries among registered agriculturists From the 43 health files data, it was found that the morbidity rates among the registered agriculturists of Roi-Et were ranked in the following order, from highest to lowest, as follows: injuries, lung diseases, WMSDs, noise-and heat-related diseases, and pesticide toxicity, in every year from 2014 to 2016. The morbidity rate of registered agriculturists was not consistent with those who identified as farmers, as recorded by the HDC, in every year from 2014 to 2016. In 2016, registered agriculturists in Roi-Et had a higher morbidity rate of all diseases and injuries than Udon Thani, except pesticide toxicity. The rates and trends increased every year in this manner, except for pesticide toxicity, which had a lower rate in Roi-Et when compared to Udon Thani. In Udon Thani, the morbidity rates from the 43 health files were found to have the same ranking as Roi-Et for every health problem, and all those rates had increasing trends from 2014 to 2016.
The results from the HDC database showed that the morbidity rate of all diseases and injuries had an increasing trend, except WMSDs in Udon Thani, which remained constant as shown in Figure 2.

Discussion
At present, the numbers of work-related diseases and injuries reported and cases among Thai farmers are likely to be lower than the true numbers because the health surveillance data of farmers or agricultural workers may not be accurate, as can be seen from the following findings and discussion.
By using registered farmers as the population for the morbidity rate calculation, which was different from using those who used a farmer occupational code as the population to calculate the rate from the HDC database, one could explain the inaccurate number of agriculturists found in healthcare service reports. The number of farmers from the HDC database  (2,204.8-2,354.3) may sometimes include duplicate recordings of farmers who visited healthcare providers more than once or sometimes be missing the occupational code identification of an agriculturist. That inaccurate number of farmers from the HDC database has led to incorrect results of morbidity rate. Therefore, for the next step of analysis, it was necessary to study registered agriculturists for accounted cases of occupational disease among agriculturists.
Those results of work-related diseases and injuries, excluding lung disease, among registered agriculturists, from the two case studies of Udon Thani and Roi-Et provinces from analysis of the HDC database and the 43 health files data, showed the same ranking of morbidity rate, which were injuries, WMSDs, noise-and heat-related diseases, and pesticide toxicity, respectively. Consequently, both Udon Thani and Roi-Et were possibly representative of national rates. For the case of lung diseases, the number of cases who had lung disease specific to the age group of between 15 to 59 years of age was analysis for the morbidity rate from HDC database, the disease was not further reported in the HDC database as occupational disease as mentioned before. It showed the highest morbidity rate from our analysis for chronic lower respiratory tract disease, or a specified lung disease among occupational diseases. However, when considering the morbidity rates (per 100,000 registered farmers) of diseases and injuries from 2014 to 2016 found in the 43 health files data, it was found that the highest morbidity rate from our analysis was not for a specified lung disease, which had not been found before with regard to occupational disease in the disease surveillance system of the Ministry of Public Health, Thailand. The high morbidity rate of occupational lung disease, which was not yet covered by the HDC database, might be explained by the fact that in disease diagnosis, it was not found to have an external cause, which would require recoding to Y96 to specify an occupational disease group. Some other diseases related to exposure to agrochemical hazards were recommended from the previous studies, e.g., Parkinson's disease 38 ; cancers of specific organs, e.g., lung, pancreas, lymphohematopoietic organs, bladder, prostate gland, 39 or thyroid 40 ; gestational diabetes mellitus (GDM) 41 or diabetes 42 ; etc. In covering those diseases to build up the surveillance system among agriculturists, some more investigation is still needed to confirm an association between each disease and a specific toxic substance used in pest control. Moreover, occupational lung disease should be considered in health surveillance of disease among farmers as per the previous report done by using a systematic review among Thai agriculturists. 43 Excluding lung disease, the three top-ranking rates across all regions and years were injuries and/or WMSDs, followed by hearing loss and diseases from heat, and pesticide toxicity, respectively. In addition, the rate of injuries was highest among all occupational diseases in Roi-Et, Udon Thani, and the nation. Udon Thani, which was chosen as a good representative province of PH Region 8, had a top-ranking rate which was the same as the nation, namely that of injuries, followed by WMSDs, noise-and heat-related diseases, and pesticide toxicity, respectively. These findings confirmed the ranking rate results found among registered crop farmers in Nong Bua Lamphu, one of the provinces in PH Region 8, Thailand; 24 however, all morbidity rates of WMSDs were more than 10 times lower than those found in other countries in Asia, 2 Canada, 3 and Europe. 4 With regard to noise-and heat-related disease, the rate found was 100 times lower compared to that of America, 5,7 and the rates of injuries and pesticide toxicity were found to be extremely low when compared to those found in Nepal 10 and China. 16 The closest morbidity rate to global rates was that of the cases of lung disease, which was more than two times lower than that of South Africa. This finding of low morbidity rates from the health database of Thai agriculturists when compared to global rates might be explained by underestimation of the morbidity rates of occupational disease and injuries.
In cases of toxicity from pesticides in 2016, the morbidity rate in Roi-Et Province was lower than previously reported among southern Roi-Et farmers during the same period, 26 which could be explained by local agricultural behaviors and the fact that pesticide toxicity report cases depended on the disease surveillance of each regional public health service. 44 Moreover, it was noteworthy that Roi-Et provincial-level case reports were very low in number compared to those of Udon Thani (Table 3).
The accuracy of the health databases varies based on the hospital level; for instance, data from hospitals at the provincial level are more accurate than those at the district level and all collected data in this study are summations of sources of data at the provincial level, which are sent to the health data centre at the national level. On the other hand, inaccuracies and actual missing data exist in Thailand, which is a major problem of diagnosing occupational diseases with a coding of Y96, and is one main problem that has been reported before. 44 However, this study used a method of matching cases with an ICD10 diagnosis code of individuals identified by national ID to cases of those who had a registered occupation of agriculturist; both types of cases can be found in databases which present the situation of occupational disease among agriculturists in Thailand.
According to the analysis of ICD-10 utilized from the 43 health files standard among registered farmers, the morbidity rate of WMSDs in Roi-Et was around double that of Udon Thani. In contrast, the morbidity rate of WMSDs in Udon Thani, according to the HDC database, was around 10 times higher than that of Roi-Et. Regarding this comparison of morbidity rates of WMSDs, namely the HDC database versus the 43 health files data records, the trends were found to be unclear. This might be explained by the fact that case reports of WMSDs are not only specific to farming work, or they have multi-factor causes. 43 Another explanation might be the wrong or missing diagnostic coding of Y96, which signifies a specified occupational cause, as reported earlier regarding the health surveillance system in Roi-Et province. 44 However, from our data of the cases extracted from the 43 files of the most recent year, 2020 45 it was confirmed that WMSDs had an increasing trend and a higher morbidity rate in Roi-Et than in Udon Thani province. Intriguingly, it was found that between the years 2014 and 2016, the province of Roi-Et had the fourth-highest tonnage of inland rice production in the country. 46 In comparison to Udon Thani, Roi-Et's in-season rice yield was approximately 2-3 times higher. 46 This study discovered from the 43 health files data that the WMSDs rate in Roi-Et was greater than that in Udon Thani province, presumably because rice farming requires more body exertion in various agricultural activities throughout the year than farming of other crops, e.g., cassava and sugar cane, which is planted more in Udon Thani.
According to HDC database analysis, Udon Thani had a higher rate than Roi-Et with regard to injuries, and diseases caused by noise and heat and pesticide toxicity; increasing rates were also observed from 2014 to 2016 among Udon Thani and Roi-Et agriculturists. Those two occupational disease groups as well as injuries were confirmed as the major health problems among agriculturists in Thailand. 43 Infectious diseases, such as leptospirosis and melioidosis, had the highest prevalence in farmers, but there was no infectious disease group shown for occupational disease in the HDC database. This was not only the case in Thailand; a previous study in Poland showed a significantly higher rate in farmers who had been exposed to repeated tick bites, 47 which corresponded to annual reports in Greece from 2004-2015. 15 Moreover, brucellosis cases have been found in farmers and livestock breeders, with a high incidence rate of 7.1 per 100,000 population. 15 From the HDC database, it was found that the ranking of the morbidity rates, in Roi-Et, in 2014, excluding that of lung disease, were as follows: morbidity rate of injuries, followed by that of WMSDs, pesticide toxicity and noise-and heatrelated diseases, respectively. However, during 2015 and 2016, the morbidity rates changed in order. Meanwhile, the ranking order of morbidity rates of Udon Thani (2014-2016) was the morbidity rate of injuries, followed by that of WMSDs, noise-and heat-related diseases and pesticide toxicity, respectively. That explains why Roi-Et was unable to provide good representative data because the province had not reported data as forecast, which contrasted with Udon Thani province in this study. The morbidity rates from the 43 files health data standard of Roi-Et province showed congruous results among the registered farmers. The highest rate was for work-related injuries, followed by WMSDs, hearing loss and heat-related disease, and pesticide toxicity, respectively. Similar results were found in Udon Thani, where rates of work-related injuries, WMSDs, and noise-and heat-related disease were similar and close to the national rates. Hence, these results reveal that the provincial data of Udon Thani can be representative of the nation. The consistency of the database is shown by the top three highest morbidity rates of disease from the previous study in crop farmers of Nongbualamphu province, which were shown to be WMSDs, noise-and heat-related disease, followed by skin irritation. 24 Moreover, the number of reported cases was very different between Udon Thani and Roi-Et, and the morbidity rate of pesticide toxicity in Roi-Et was dramatically lower than that of the previous report from 2016-2018 among registered farmers in Sakon Nakhon, a province in PH Region 8. 25 Those cases mentioning skin disease were suspect due to a combined effect in the group of pesticide toxicity, which is that of a reported acute symptom (irritant), and confirmed the previous studies. 25,26 Regarding agricultural productivity in the study area, the Office of Agricultural Economics reported that cassava, rubber, and rice had the highest productivity between 2016 and 2019, 46 as found in the case characteristics of this study. That could support results concerning the significant increase in morbidity rates of all occupational diseases and injuries in Roi-Et from 2014 to 2016; this increase was also found in Udon Thani, particularly in regard to pesticide toxicity. Pesticide toxicity has been shown to be a potential disease of the agricultural sector of Thailand, and as discussed, that problem was exacerbated by the trend of increasing pesticide imports, particularly herbicide imports, during the 5-year period from 2013 to 2017. 48 In comparison to Roi-Et province, Udon Thani has almost five times the production areas of cassava and Para rubber. 46 This study reveals that Udon Thani has a higher increasing trend of morbidity rate for pesticide toxicity than Roi-Et province. Hence, it is clear that herbicides are being used continuously to treat the rubber and cassava while it is being harvested.
These disputations with strong evidence suggest health surveillance of occupational disease among agriculture workers (informal workers) or Thai farmers by 1) realizing the diagnosis with a coding record of Y96, 2) improving the recording method of occupation to be used practically for agriculturists, and 3) promoting agriculture workers to register on the health database. It was more than interesting that some have mentioned other diseases should be added to ICD-10 for occupational disease, i.e., lung disease, skin disease and infectious disease, for the occupational surveillance system.

Conclusions
The morbidity-rate ranking of work-related diseases and injuries, excluding lung diseases, among agriculturists in the upper northeast of Thailand was injuries, WMSDs, noise-and heat-related disease, and pesticide toxicity, respectively. Lung disease, which was not reported as an occupational disease of the HDC database, had the highest morbidity rate among occupational diseases in Roi-Et (2,781.7 per 100,000 farmers) and Udon Thani (2,278.6 per 100,000 farmers) from the 43 health files data. The national morbidity-rate ranking in 2016 (per 100,000 farmers) from the HDC database was injuries (601.0), work-related musculoskeletal disorders (WMSDs) (354.3), noise-and heat-related diseases (104.4), and pesticide toxicity (14.6), respectively. Those morbidity rates of Roi-Et and Udon Thani provinces were representative of the morbidity ranking of diseases of the nation. When comparing the morbidity rates of diseases of farmers from the 43 health files data to those of the HDC database, the closest rates were found in the pesticide toxicity of Roi-Et (26.5 per 100,000 farmers) and Udon Thani (27.7 per 100,000 farmers). The rates of WMSDs in Roi-Et (754.3 per 100,000 farmers) and Udon thani (479.1) from the 43 health files data were more than two times higher than those rates from the HDC database (Roi-Et: 46.2 and Udon Thani: 272.4 per 100,000 farmers), which was like those of noise-and heat-related diseases as well as that of injuries. All work-related diseases and injuries of the 43 health files data and almost all of those of the HDC data had increasing trends from 2014 to 2016. The number of farmers in the HDC database did not consistently reflect the number of registered farmers in the agricultural database. Situations of work-related diseases and injuries discovered among registered farmers reflect the health problems of Thai agriculturists, and the underestimation in the reported disease rate in the health database is explained by big data analysis, which showed that work-related cases with an identifying code of Y96 had rarely been reported among agriculturists. Therefore, recording an occupational cause when visiting health service providers should be promoted in the future, and Thai agriculturists should be supported in registration with healthcare databases for surveillance of various types of occupational diseases and injuries, including lung, skin and infectious diseases.

Underlying data
The underlying data for this study is secondary data collected by the Health Data Centre and agricultural registries and is restricted due to ethical and data protection considerations. Under the ethical approval from the Human Research Ethics Committee of Khon Kaen University (No. HE592154), participants' data was not allowed to be shared and the secondary data of health dataset and registered farmer dataset were utilized under restriction. The individual information or further reused inclusion data is not permitted under the ethically approved conditions for human research. This was confirmed under the data protection of human/participant privacy as no individuals gave explicit written consent that their identifiable data can be made publicly available. For more information, this research had no personal identified data for accessible because of the ethical issue that does not allow us to hold the personal data of each participant and cannot be shared any data after closing research.
The raw dataset collected under permission access and restriction use of ICD-10 codes are the 43 health files lists of Roi-Et and Udon Thani provinces in the period 2014 to 2016: file 10diagnosis (opd), file 15diagnosis (ipd), and file 19surveillance and the lists of registered agriculturists in Roi-Et and Udon Thani provinces during 2014-2016. To request access to the underlying data, please email the primary investigator (Sunisa Chaiklieng; csunis@kku.ac.th as the corresponding author. Applications will be reviewed on a case-by-case basis.
The intermediary data that can be de-identified without compromising anonymity are all presented already in the results and underlying data (TABLE A,  The project contains the following underlying data: •  Figure 1; Table B -data for Figure 2 that is related to HDC database; Table C -data for Figure 2 that is related to 43 Health files; Results -should show key findings in numbers 2.
Registered farmers may not be the true population at risk, as discussed in Conclusions; moreover, this bias may not be equal in the 2 provinces 3.
Other potential biases are -discrepancy of diagnoses and codings of the 2 provinces and across the 3-year period, different quality of diagnoses and codings of different levels of 4.

hospitals, etc. -these should be addressed and discussed
Be aware of the difference between 3-year report/statistics and 3 consecutive-year report/statistics 5.
Neurotic hearing loss -I do not agree to use this term. There are sensorineural, conductive and mixed hearing losses.

6.
Regarding each individual province, Roi-Et had an increasing trend for all diseases, for both data from the 43 health files data and data from the HDC, but a disease of Udon Thani had a clearly decreasing trend, --I do not understand and am not sure which disease is " a disease of Udon Thani".

If applicable, is the statistical analysis and its interpretation appropriate? Partly
Are all the source data underlying the results available to ensure full reproducibility? Yes

Are the conclusions drawn adequately supported by the results? Yes
Competing Interests: No competing interests were disclosed.

Reviewer Expertise:
Occupational and environmental medicine/health, preventive medicine, public health I confirm that I have read this submission and believe that I have an appropriate level of expertise to confirm that it is of an acceptable scientific standard, however I have significant reservations, as outlined above.

Dear Prof. Dr. Pornchai Sithisarankul
Thank you for your valuable comments, we have improved paper following your suggestions, as the revised version. The clarification for each question is following.

Background -work-related diseases and injuries -so the result should show or address
Answer: This results and discussion of paper was improved to address injuries data as suggestion.

Results -should show key findings in numbers
Answer: The results was improved to show key findings in numbers as suggestion.
3. Registered farmers may not be the true population at risk, as discussed in conclusions; moreover, this bias may not be equal in the 2 provinces Answer: The most recent agricultural census, conducted every ten years, was conducted in 2013. In each province of the nation, data collection will take place simultaneously by sending around 20,000 officers, 17,000 counting personnel/employees, 3,000 academic staff, and farmland holders/heads of each household to conduct interviews to compile fundamental data about agricultural operations. These details will be utilized to choose appropriate agricultural properties and conduct a meeting with qualified details (enumeration). By that reason, the registered farmers in Thailand are strongly representative agricultural farmers by an occupation, resulting on occupational hazards exposure and the true population at risk for this study, as we have done in the previous studies and method [24] [25] [28]. In addition, among the registered farmers, the real strong point is the linking of the identified cases to the full farmers population approved by the Provincial Agriculture Office agriculture database, that to show the magnitude of risks of the agricultural farmers.
Moreover, the analysis in this study was accurate and have reliability by matching personality with the national ID to the health records of 43 files based on identifiers of individuals to another source of the registered farmers for Bigdata analysis. This method was done for previous studies [24,[25][26] which could show the risks of the Thai agricultural farmers on occupational diseases, particularly WMSDs, heat related illness or pesticide toxicity, that highly risks in agriculture were under-estimated from the previous report accounted 2-10 times of health database as in this study report. Any bias which might not be equal in the 2 provinces, we mentioned in the discussion to this serious and lack or difficulty to get support of recording occupational disease may have caused non-reporting as in the discussion. 4. Other potential biases are -discrepancy of diagnoses and codings of the 2 provinces and across the 3-year period, different quality of diagnoses and codings of different levels of hospitals, etc. these should be addressed and discussed.
Answer: The discussion was added for those issues, it was discussed that the accuracy of the health databases varies on the hospital level; for instance, the provincial level is more accurate than the district level and all collecting data in this study are summation at the provincial level sources of data, where sending data to health data center of a country level.
Whereas inaccuracies and actual missing data exist in Thailand is the major problem of diagnosing occupational diseases coding of Y96, which is one main problem as we discussed (Discussion section). However, this study method of analysis, we matched case of ICD10 diagnosis code to the registered agriculturist occupation (more by national ID. Both databases are the best representative for presentation the situation of Occupational diseases among agriculturists in Thailand.
The only potential bias could be by quality of diagnoses and coding of different levels of hospitals which was addressed in discussion section according to the forecast which might occur and missing diagnostic coding in case of WMSDs, moreover, pesticide toxicity report cases depended on the disease surveillance of each regional public health service. 44 5. Be aware of the difference between 3-year report/statistics and 3 consecutive-year report/statistics Answer: We aware of the difference, all parts in this paper and underlying data were rechecked and replaced with the correct term to present in this paper, for example, from 2014 to 2016, in 2016, thank you for this recommendation.
6. Neurotic hearing loss -I do not agree to use this term. There are sensorineural, conductive and mixed hearing losses. For a global audience, the first paragraph is less relevant than the 2nd. I would suggest moving it down to a Thailand specific focus before the HDC paragraph.
In the literature review paragraph, you might start by citing some of the reviews of the range of agriculture and health in general before going into each disease-disorder-injury type.
Re the HDC, need some more background about its role in surveillance or health data in Thailand.
A case-control study does not usually result in prevalence estimates. Clarify how this might provide such an estimate.
Good to have methods of identification of the above. Seems like preliminary work to this paper. Might it be better as an early part of methods?
This work is clearly important for Thailand. How might it be important globally, in other MICs for example? How does it relate to the Global Burden of Disease work e.g. on low back pain? Something more generalizable as a scientific objective is needed. For example, it seems that there are two: one around work-related injury and disease surveillance methods, in which the paper makes a contribution; and one around differential burdens in different settings and exploration of reasons.
Great use of a registered farmer -farmworker data set. Any sense what the breakdown would be i.e. proportions with different access rights to land? employers or employees?
Not sure what you mean by "more real" agriculture? what is "less real" agriculture? How did you decide on the occupational codes in table 1?
In some jurisdictions, farmer suicides from pesticides are considered work-related, as reflect the difficult situation of many farmers, with cash and debt squeezes which force them into suicide as one of few options. And pesticides are associated with depression. Please justify your exclusion e.g. more of a narrow safety view of exposure.
Can you provide some sense of the completeness of coding of health conditions etc. This is often the challenge with regards to occupational disease and occupational injury occurrence.
Data analysis, last two sentences of first paragraph should be edited together. Formula "of" or "in" rather than "at" area of interest.
I found it confusing to have the number of registered farmers in the same table as number of HDC farmers. the latter is the denominator for table 3 and Figure 1, while the former is the denominator for subsequent analysis shown in table 4, where this denominator of registered farmers should be included.
Not sure that the interprovincial comparisons are as important globally as they may be internally for national and provincial policy makers. Less emphasis perhaps, unless going to speak to the heterogeneity across regions within countries globally, and some of the reasons for such heterogeneity e.g. industrial sector profiles, meteorological conditions, etc.
Agree that under-reporting common in surveillance. More reference to this methodological literature, the strengths and weaknesses of different systems, particularly for work-related injuries and diseases is needed.
What factors might explain the increasing morbidity trends? Is more awareness, education and diagnosis occurring? more exposure? Need some exploration of reasons. Any data sources that might inform this discussion?
Currently much repetition of results here, without linkage to the global literature. How do the rankings here compare with other country rankings of morbidity among agriculturalists? and why, do you think.
Why do you think health service providers do not record occupation in Thailand? How does it compare globally?
Really appreciate the data availability statement and your work in this regard. The OSF tables are more other versions of the results, however.

Are sufficient details of methods and analysis provided to allow replication by others? Partly
If applicable, is the statistical analysis and its interpretation appropriate? Yes 6. This work is clearly important for Thailand. How might it be important globally, in other MICs for example? How does it relate to the Global Burden of Disease work e.g. on low back pain? Something more generalizable as a scientific objective is needed. For example, it seems that there are two: one around work-related injury and disease surveillance methods, in which the paper makes a contribution; and one around differential burden in different settings and exploration of reasons.
Answer: From the big data analysis from the 43 health files is clearly identified cases of occupational disease to estimate the morbidity rate by extracted ICD-10 code of occupational disease, for example, code M545 of the registered farmers holding cultivating activity who was matching national ID to the accessed healthcare farmers. The previous study among crop farmers [reference 24] could show the number of low back pain cases of 4,814 cases in 2016 that reflected the highest morbidity rate of WMSDs (21.73%) among all occupational diseases in registered farmers. The study discovered also lung diseases, which was not reported in HDC database as the occupational disease.
The real strong point of the paper is the linking the identified cases to the full farmers population and the outcome is showing the magnitude of risks of the agricultural farmers, which is a large part of the workforce in not only in Thailand but throughout globally. Even in highly developed countries hazards in agriculture are usually under-estimated, and negative outcomes not identified due to compensations mechanisms as we mentioned in the discussion continued from the background of Thai labor protection and compensation among agriculturists who are informal workers in Thailand. This study used two provinces of Roi-Et and Udon Thani which are two of the provinces among the top five provinces engaged in cultivating activities in the Northeast of Thailand and the major of planting area and numbers of farmers in Thailand. The population of agriculturists in those two provinces represented Thai agriculturist as the nature of the impact from agricultural work which was found to be similar across from other areas of Thailand because the main product of all regions in Thailand was rice. Therefore, this paper could make a contribution to the disease surveillance method of work-related injury and diseases as we improved written more focusing in the abstract and discussion and the conclusion. 7. Great use of a registered farmer -farmworker data set. Any sense what the breakdown would be i.e. proportions with different access rights to land? employers or employees?
Answer: The most recent agricultural census, conducted every ten years, was conducted in 2013. In each province of the nation, data collection will take place simultaneously by sending around 20,000 officers, 17,000 counting personnel/employees,3,000 academic staff, and farmland holders/heads of each household to conduct interviews to compile fundamental data about agricultural operations. These details will be utilized to choose appropriate agricultural properties and conduct a meeting with qualified details (enumeration). That strongly confirm that the registered farmers in Thailand are representative agricultural farmers as occupation, resulting the most true population at risk on occupational hazards exposure for this study. Roi-Et and Udon Thani which are two of the provinces among the top five provinces engaged in cultivating activities in the Northeast of Thailand were selected as representative provinces. The major of planting area and numbers of farmers in Thailand located in Northeast of Thailand, and the main product of all regions in Thailand was rice as presented in Roi-et and Udon Thani. For accessing of healthcare, this study, by utilizing the 43 health files data of the ICD-10 code of occupational cases and injuries during 2014 to 2016, it was found that there were 53,794 total cases (34.82%) of registered farmers who had accessed healthcare in Udon Thani, and 77,438 total cases (37.32%) in Roi-Et, that was not much different of those provinces. For the cases visited of healthcare as shown as additional results, the characteristics of those cases of registered farmers in Roi-Et and Udon Thani provinces, the largest proportion of cases in Roi-Et were working on rice plantations (98.26%), as well as in Udon thani (89.1%). Those showed only in numbers of the difference because of cultivating activities were different during off-rice season as we discussed.
Thai's agribusiness is clear content with contact farming laws which registered farmers are taken care of by government benefits such as price insurance. This study year have no statistics of agribusinesses number because agricultural census is providing the survey of agricultural holdings and will be report in 2023. The updated agricultural census was reported about statistics of agribusinesses which they provided by survey the registered farmer. At the present, there are no contact farming entrepreneur registered with the government in Roi-Et and Udon Thani. 8. Not sure what you mean by "more real" agriculture? what is "less real" agriculture? How did you decide on the occupational codes in table 1?
Answer: Roi-Et and Udon Thani are two of the provinces among the top five provinces engaged in cultivating activities in the Northeast of Thailand (comprising more than 97.0% of agricultural holdings) as explained before, therefore, the word of "more real" was deleted to be clear from the design of this study.
For the occupational codes in table 1 was the recorded code of occupation in HDC database as we already explained in method section; All codes of farmers presented in Table 1 are the cultivating famers or involved activities in cultivated farming as the reference code of farmers from the Health Data Center (HDC) ( https://hdcservice.moph.go.th/hdc/main/index.php) and confirmed by the IT personal responsible of coding and the agriculturist expertise. 9. In some jurisdictions, farmer suicides from pesticides are considered work-related, as reflect the difficult situation of many farmers, with cash and debt squeezes which force them into suicide as one of few options. And pesticides are associated with depression. Please justify your exclusion e.g. more of a narrow safety view of exposure.
Answer: This study aware of the suicide from pesticide and we already excluded the cases not be accounted to the morbidity rate of pesticide toxicity as shown at notes in Table 2. In addition, the case of patient death during the study year which was accessed from file number 3 of the 43 health files were excluded from the visited cases in this study.
10. Can you provide some sense of the completeness of coding of health conditions etc. This is often the challenge with regards to occupational disease and occupational injury occurrence.
11. Data analysis, last two sentences of first paragraph should be edited together. Formula "of" or "in" rather than "at" area of interest.
Answer: It was corrected as follow;

Number of cases reported among registered agriculturists or farmers visited services in an area of interestTotal number of farmers (registered agriculturists or farmers visited services) in an area of interest
12. I found it confusing to have the number of registered farmers in the same table as number of HDC farmers. the latter is the denominator for table 3 and Figure 1, while the former is the denominator for subsequent analysis shown in table 4, where this denominator of registered farmers should be included.
Answer: The numbers of registered farmers of Roi-Et and Udon-Thani referred to Table 3 were corrected as presented for table 4. 13. Not sure that the interprovincial comparisons are as important globally as they may be internally for national and provincial policy makers. Less emphasis perhaps, unless going to speak to the heterogeneity across regions within countries globally, and some of the reasons for such heterogeneity e.g., industrial sector profiles, meteorological conditions, etc.
Answer: Thank you for this point and that is agree with the design of this study and the results, we have improved the abstract and conclusion to be less emphasis of interprovincial comparison, but representative the rate or trend of morbidity rate of health problems of Thai agriculturist was presented, as shown in the Abstract and Conclusion. For the discussion in the morbidity rate of different, it is mostly pointed for WMSDs which showed the contradictory rate and ranking orders. The heterogeneity across regions within countries, the agricultural productivity while off-rice seasoning was more discussed also for pesticide toxicity.
14. Agree that under-reporting common in surveillance. More reference to this methodological literature, the strengths and weaknesses of different systems, particularly for work-related injuries and diseases is needed.
Answer: We confirm the previous studies (24-25) and the preliminary results of the southern Roi-Et (26), of the under-reporting cases from the surveillance system of occupational diseases (44) as discussed. This study showed one pational disease in HDC database, by the method of extracted ICD-10 coded of J40-J47, could be done from the 43 health files data matching to the agriculturists. The strength of Occupational diseases surveillance in the HDC database is the platform cases recorded from the Universal healthcare service which is usually covered for Thai population and farmers in Thailand. However, there were some weakness, 1) at time of this study, is that the recording of occupational disease was not mandatory but depended on the decision of hospital management and resources. Workrelated disease effects on examination and treatment were not confirmed, for example, in the case of silicosis; it has symptoms similar to tuberculosis (TB), therefore, the report might be missing because it was not confirmed by silica dust exposure in the workplace from stone and sand, or mining or sculpture, or working in the cement industry. Moreover, for example lung diseases, infectious diseases, had not been included in the HDC database of Thailand as types of occupational disease. 2) The accuracy of the health databases varies based on the hospital level; for instance, data from hospitals at the provincial level are more accurate than those at the district level and all collected data in this study are summations of sources of data at the provincial level, which are sent to the health data centre at the national level.
3) The occupational codes for farmers as shown in Table 1 in HDC database seems to be complicated for recording during routine work for healthcare providers. For this weakness, the second method to use the number of registered agriculturists can be solve the more strength method.
The strength of the method utilizing the ICD-10 from the 43 health database are the following; 1) the morbidity rate of lung disease estimation which can be provide among registered farmers by matching personal ID of cases to those farmers in agricultural database.
2) As we found the morbidity rate for WMSDs was the highest concerned diseased among all diseases, WMSDs ICD-10 codes in Table 2 were categorised in the health surveillance system, but they only allow for the knowledge of disease rates and cannot be used to identify the organs that are affected in WMSDs as we have shown from the recent study report [47; the WMSDs: codes of diagnosis were M00-M99, M542, M545, M548, M60, M624, M653, M792, M796, G560], 3) limitation of diagnosis which is depended on the accessed healthcare service level. That is equal from the HDC database and the 43 health files method. 3) Another weakness to be concerned from the disease surveillance system of the use of ICD-10 code with occupational causes, for example, pesticide toxicity, a previous study in Roi-Et province found that ICD-10 identification and reports on the disease surveillance system with external code Y96 (work-related condition) were missing, as were reports in severe cases of pesticide poisoning with T-toxicity coding (44). Therefore, pesticide toxicity was shown to be underestimated in regard to occupation, particularly that of cultivating farmers, and the analysis utilised ICD-10 from the 43 health file data could include more cases without code Y-96 to estimate the rate.
Thailand's public health policy should include ICD-10 code L (skin disease) and code J (lung diseases) to identify the acute and chronic effects of pesticide toxicity among applicators, as recommended in our previous study of cultivating farmers in Sakon Nakhon province (26). By that considered ICD-10 code L, the stronger method of 43 health files data can provide manually extract code ICD-10 of the cases of the registered farmers.
15. What factors might explain the increasing morbidity trends? Is more awareness, education and diagnosis occurring? more exposure? Need some exploration of reasons. Any data sources that might inform this discussion?
Answer: From both data source, and we provided more data as shown in underlying data, Table D, and the discussion continued to the conclusions, the trend of increasing rate of diseases was presented. It can be explained from the main factors of health hazards exposure from cultivation activity and types of plantations affecting the different diseases rates of the two province-case study as discussed for pesticides toxicity and WMSDs increasing trend in the Discussion as the follow.
"Regarding agricultural productivity in the study area, the Office of Agricultural Economics reported that cassava, rubber, and rice had the highest productivity between 2016 and 2019 (48), as found in the case characteristics of this study. That could support results concerning the significant increase in morbidity rates of all occupational diseases and injuries in Roi-Et from 2014 to 2016; this increase was also found in Udon Thani, particularly in regard to pesticide toxicity. Pesticide toxicity has been shown to be a potential disease of the agricultural sector of Thailand, and as discussed, that problem was exacerbated by the trend of increasing pesticide imports, particularly herbicide imports, during the 5-year period from 2013 to 2017 (49)". …."However, from our data of the cases extracted from the 43 files of the most recent year, 2020 [47], it was confirmed that WMSDs had an increasing trend and a higher morbidity rate in Roi-Et than in Udon Thani province. Intriguingly, it was found that between the years 2014 and 2016, the province of Roi-Et had the fourth-highest tonnage of inland rice production in the country (48). In comparison to Udon Thani, Roi-Et's in-season rice yield was approximately 2-3 times higher (50). This study discovered from the 43 health files data that the WMSDs rate in Roi-Et was greater than that in Udon Thani province, presumably because rice farming requires more body exertion in various agricultural activities throughout the year than farming of other crops, e.g., cassava and sugar cane, which is planted more in Udon Thani." 16. Currently much repetition of results here, without linkage to the global literature. How do the rankings here compare with other country rankings of morbidity among agriculturalists?
Answer: The discussion was improved by linking to the global literature and compared with the ranking of morbidity among agriculturist as in section of discussion as the follow; "For noise and heat related disease, the finding rate was 100 times lower compared to of the America (5, 7), which is similarly lower direction found in the rates of injuries and pesticide toxicity when compared to those found in Napal (10) and China (16). The closest rate to globally was that for the cases of lung disease which is lower than 2 times lower than in the south Africa. This finding too low morbidity rate from the health database of Thai agriculturists when comparing to those global rates might be explain the underestimation of the morbidity of the occupational disease and injuries. As we have a real strong point of the paper from the method from 43 health file big data analysis linking the identified cases to the full farmers population and the outcome is very showing the magnitude of risks of the agricultural farmers, which is a large part of the workforce not only in Thailand but throughout Asia and globally.
17. Why do you think health service providers do not record occupation in Thailand? How does it compare globally?
Answer: By HDC, the occupation code is available for case visited for healthcare service as we use the code of farmers involved plantation activities during the period of in this study, however, the many codes available to be selected to be represented farmers occupation as shown in Table 1 which is complicated for the healthcare providers to add in the record. There are also codes of unemployed, general employed, underling person, or no occupation, that could be selected for this filed, which made missing/misclassification in database. The previous study of pesticide poisoning reports among showed incomplete records of occupational cases in 45.14% of all cases, and misclassification case report of agriculturists as occupation code was found in 37.03% of cases in Skon Nakhorn from our previous case-control study [Chaiklieng, S.; Khamjantarat, P.; Suggaravetsiri, P. Factors as sociated with disease from pesticides toxicity among planting farmers in Sakon Nakhon Province. In: Proceedings of The 9 th National Conference in Toxicology (NCT9), Update in toxicology for safe life. 25-26 September 2019, 109-21]. Moreover, there were key concerns regarding public health personnel being responsible for multiple tasks, without any understanding of occupational disease codes, and never having received training for occupational disease recording [44].
As you agreed of under-reporting, and I am sure that even in highly developed countries, hazards and risks in agriculture are usually under-estimated, but negative outcomes not identified due to many reasons, i.e., lack of legal, control and compensations mechanisms. By those reasons, almost no recently update report to compared since 2004 after one study mentioning to that occupational history which was frequently neglect occupational history-taking by the physicians, and the physician had completed an occupational history-taking from only 27.8% of patients (Politi et al. Occupational Medical History Taking. JOEM 2004;46:550-5), which is very low and also lower rate compared to our previous finding in Thailand [44].
18. Really appreciate the data availability statement and your work in this regard. The OSF tables are more other versions of the results, however.