Meta-Analysis of the Disease Spectrum of Hospitalized Crew-Members

Background: The disease spectrum of hospitalized crew-members could give good service to the support of health management. However, the disease spectrum of Chinese crew-members is still obscure. Methods: Four different database were searched with keywords. Papers were included and excluded using given rules. Data such as publishing year, investigation period, subject resources, total number of subjects, name of diagnosis, proportion of diseases were extracted for investigation. Then the investigation year(before 2010 or after 2011) was applied for sub-analysis in further research. Results: 24 results of disease spectrum of crew-members from numbers of hospitals were meta-analyzed. The rst common disease was cervical and lumbar spine disorder which occupied 10% of diseases for hospitalization. And the other diseases followed it were hyperlipidemia(8%), cervical spine disease(7%), neurasthenia(6%), hypertension(6%), gastritis and duodenitis(6%), headache(6%), vegetative nerve functional disturbance(6%), upper respiratory infection(6%), fatty liver(5%), gastric and duodenal ulcers(4%), arrhythmia(4%), ground syncope(3%), urinary calculus(3%) and poor acceleration tolerance(2%), respectively. The spectrum of diseases was sub-analyzed by investigation year for further exploration. The proportion of cervical and lumbar spine disorder increased dramatically in the group of after 2011. Conclusion: Our work gave evidenced-based support for disease spectrum of crew-members, pointed research directions for treatment and prevention of crew-members’ common diseases, supported health strategy constitution of crew-members and guided well allocation of medical resources.

exploration. The proportion of cervical and lumbar spine disorder increased dramatically in the group of after 2011.
Conclusion: Our work gave evidenced-based support for disease spectrum of crew-members, pointed research directions for treatment and prevention of crew-members' common diseases, supported health strategy constitution of crew-members and guided well allocation of medical resources.

Background
Crew-members are groups of people working in the sky. They work in specially tough environment, frequently exposed to low pressure [1] , chemicals [2] , acceleration [3] and so on. This makes crew-members at great risk for diseases. In one word, crew-members are taking high risks at work.
Getting hold by diseases and being detained in hospital would stop them from working in the air. Once there is a problem with the health of crew-members, the aeronautical training and duty will be seriously interrupted [4] . Therefore, the support for health management of crew-members is profoundly crucial.
The disease spectrum of hospitalized crew-members could support well to health management of crewmembers [5] . The disease spectrum is the summarized results of common diseases of crew-members. It could recognize the frequently happened diseases that stop them from normal work and training. It gives a brief descriptions of diseases causing crew-members' hospitalization [6] . It could also be sub-analyzed with different factors, such as time, age for further investigation [6] . It also draw attention for aviation doctor and scholars on typical diseases, especially those at the top of the spectrum. In this way, the specialists could focus on the reason of variation of disease spectrum, also could come up with effective strategy of treatment and prevention. Therefore, the disease spectrum of crew-members could achieve better health services for crew-members.
However, the disease spectrum of Chinese crew-members is still obscure. Zhang et al concluded that cervical/lumbar spinal disease and hypertension are top 2 categories for hospitalized crew-members, up to 9.30% and 7.97% respectively [7] . But Yang et al, through collection of diseases causing hospitalization of pilots, realized the top two ranking diseases are cervical/lumbar spinal disease and hyperlipidemia, up to 8.03% and 7.05% respectively. And the percentage accounted for hypertension in their disease spectrum only take 1.86% [6] . These results lacking consensus in proportion of diseases of hospitalized crew-members.
The reason for getting hard to draw consensus of disease spectrum may as follows. First, the overall number of crew-members is small, let alone the small numbers of crew-members enrolled in the investigation of disease spectrum. Second, the reports of disease spectrum comes from one hospital.
This lacks integrated analysis with numbers of centers which might be another reason for differences of spectrum.
Therefore, our paper collected the published articles from multi-center that related to Chinese crewmembers spectrum, extracted and synthesized the data from these articles which involved huge number of subjects, in order to give a robust evidence of the spectrum of Chinese inpatient crew-members. We came a conclusion for disease spectrum of Chinese crew-members. The frequently happened diseases were presenting in our disease spectrum of Chinese crew-members. We further sub-analyzed the spectrum of diseases by investigation year for advanced exploration. Our work gave evidenced-based support for disease spectrum of Chinese crew-members, pointed research directions for treatment and prevention of crew-members' common diseases, supported health management of Chinese crewmembers and guided well for allocation of medical resources.

Data sources and search strategy
Two investigators(Zhouheng Ye and Tianyi Zhang) conducted systematic literature review using dataset. The database included China national knowledge infrastructure (CNKI)(Published from 1990 to 2020), Wangfang Data(Published from 1990 to 2020), VIP Database for Chinese Technical Periodicals (VIP) (Published from 1990 to 2020)and PubMed(Published from 1990 to 2020). Papers related with description of disease spectrum of Chinese crew-members were searched. While the search formula is varied according to different database, the search keywords is same in each database, named as "aircrew " and "inpatient or hospitalization" and "disease spectrum". The Chinese version of search keywords was used in Chinese Database. The same paper was merged after reading essential information, such as title, author and publication year, of each paper. Then the full copy of the ltered paper was downloaded for further review.

Inclusion And Exclusion Criteria
The inclusion and exclusion criteria are as follows. Inclusion criteria: Subjects are crew-members from mainland of China; The research style is cross-sectional study; The subjects' hospitalization is caused by diseases; The diagnosis of disease is clear, speci ed as organ + disease; Data resources is wellde ned. Also clear are the spectrum period, total number of subjects and percentage for typical disease.
The diseases with over 6 reports of proportion were quali ed for inclusion. Exclusion criteria: The paper is reported by sanatorium and lack speci c diseases; Diagnosis is de ned in the way of systematic diagnosis, such as digestive disease or in the way of department + disease, such as orthopedic disease, endocrinology disease and so on; The total number of subjects is missing or the proportion of disease is missing; Research focus only on specialized elds and results of disease spectrum showed only on specialized elds. Study selection and application of inclusion criteria were carried out independently by the two investigators who conducted the literature search(Zhouheng Ye and Tianyi Zhang).

Data Extraction
Data extraction was independently performed by two investigators(Zhouheng Ye and Tianyi Zhang). The following information was extracted from each eligible study: author, year, study period, subject resources, total number of subjects, name of diagnosis, proportion of diseases.

Search results and study characteristics
The complete selection process is as follows (Fig. 1). 206 records were retrieved by our search strategy. We excluded 165 articles after reading the titles and abstracts, and retained 41 articles for further evaluation by reading the full articles. Finally, we selected 24 full-text articles about hospitalization disease spectrum of Chinese crew-members. The sample sizes reached up to 26983 within the analysis.

Sub-analysis
We divided included papers by investigation year. The sub-analysis of investigation year applied to 14 kinds diseases out of all common diseases. The proportion of disease in sub-group was meta-analyzed using same way above.

Statistical analysis
We calculated the proportion of disease spectrum of hospitalized crew-members based on the proportion and 95% con dence intervals published in each study. Meta-analyses were conducted in R version 4.0.3, using the MetaProp command in the "meta" package [8] . Pooled proportions were estimated using random effects models. Heterogeneity between studies was evaluated using the I2 statistic (p-values for heterogeneity). All statistical tests were 2-sided, and a P-value < 0.05 was considered to be statistically signi cant. Table 1 showed general characteristics of the 24 studies included in the analysis. The year of publication ranged from 1996 to 2019. The investigation duration was from 1958 to 2018. There are total 26983 subjects involved in meta-analysis. Fifteen kinds of diseases were involved in the investigation. Cervical and lumbar spine disorder included 23 research papers for investigation which was the most within all diseases. Neurasthenia included the least for investigation, only 6 papers. Table 2 showed the proportion of diseases that is commonly happened in the hospitalized population of Chinese crew-members using meta-analysis. The rst disease was cervical and lumbar spine disorder which occupied 10% of disease for hospitalization. And the other disease followed it was hyperlipidemia(8%), cervical spine disease(7%), neurasthenia(6%), hypertension(6%), gastritis and duodenitis(6%), headache(6%), vegetative nerve functional disturbance(6%), upper respiratory infection(6%), fatty liver(5%), gastric and duodenal ulcers(4%), arrhythmia(4%), ground syncope(3%), urinary calculus(3%) and poor acceleration tolerance(2%) respectively. The cervical and lumbar spine disorder was the leading disease with pooled proportion of 10%, 95%CI 8-13%. There were 2026 subjects out of 23 research records involved in the investigation. The p value of heterogeneity was < 0.05 while the p value of egger's test was over 0.05. Hyperlipidemia ranged second place of the results with pooled proportion of 8% (95%CI 5-11%). The third disease was cervical spine disease with pooled proportion of 7%(95%CI 4-12%). Neurasthenia was the fourth common disease with pooled proportion of 6%(95%CI 3-9%). This proportion was same with hypertension, gastric and duodenitis, headaches, vegetable nerve functional disturbance and infection of the upper respiratory tract. The p value of the disease above was all < 0.05 while the p value of egger's test was over 0.05. The poor acceleration was the last disease in table 2 with pooled proportion of 2%(95%CI 2-3%). The other included disease, such as fatty liver, gastric and duodenal ulcers, arrhythmia, ground syncope and urinary calculus, was below 6%, ranging from 3-5%(table 2). Figure 2 representatively showed the forest plots of cervical and lumbar spine disorder. The blue marker demonstrated the proportion extracted from each included records while the red diamond showed the pooled proportion meta-analyzed by metaprop function.

Results
Sub-group analysis was applied according to different investigation year( Table 3). The 2010 was set as the dividing line for sub-analysis of each disease. Cervical and lumbar spine disorder showed proportion of 10% in group of before 2010 while a remarkable increase to 15%(95%CI 9-26%) was found in group of after 2011. The cervical spine disease reported proportion of 5%(95%CI 3-8%) in group of before 2010 while in group of after 2011 the pooled proportion increased to 12%(95%CI 4-38%). The egger's test for both disease was over 0.05 while p value of heterogeneity was < 0.01. The vegetative nerve functional disturbance had higher proportion in the group of before 2010(9%, 95%CI 6-11%) compared with the group of after 2011(3%, 95%CI 2-5%). The fatty liver showed pooled proportion of 2%(95%CI 1-4%) which raised to 6%(95%CI 5-8%) in the group of after 2011. The arrhythmia showed pooled proportion of 3%(95%CI 2-5%) which grew to 6%(95%CI 3-9%) in the group of after 2011.

Discussion
The disease spectrum of the Chinese crew-members can describe the common disease of the crewmembers that is crucial for health management of Chinese crew-members [9] . However, the disease spectrum is not well documented and lacks consistence between each research paper [10] . Thus, with poor health management causing by inde nite disease spectrum, the crew-members cannot ful ll the job working in the air. And there is no way to guide properly for health support for aviation doctor. Therefore, the necessary to recognize the most common disease for the crew-members is deeply accepted.
The rst common disease was cervical and lumbar spine disorder. There is a consistency that cervical and lumbar was the rst common disease in spectrum of Chinese crew-members. Yang et al supported this idea with their results [6] . The paper of Yang's presented that the rst three common disease in the investigation was cervical and lumbar spine disease, hyperlipidemia and fatty liver. The proportion of the disease was 8.03% which means 82 got the diseases out of 1021 subjects. This prevalence is same with our results, 10%(8%-13%). Our previous work had higher results that the two most frequently happened diseases were cervical spondylosis and lumbar disease [11] . The proportion of these two diseases were 29% and 22% respectfully. Xiuming Wang and colleagues reported the rst common disease was lumbar and cervical disease [12] . Our results reported proportion with 10%. It is between the result of Jiaxing Zhou et al and Xiuming Wang.
There might be three reasons for crew-members getting a disc herniation. First, crew-members have to keep a sitting statue for long time, especially for transport plane. Second, the helmet may burden the pressure hold by cervical spine, especially when there is high acceleration [13] . Third, the plus-minus acceleration may introduce more pressure on spine to ght against the acceleration [14] .
Our sub-analyzed results of cervical and lumbar spine disorder of investigation year supported the third hypothesis. We did subgroup analysis for group of year before 2010 and group of year after 2011. There is an increase of acceleration for Chinese crew-members after 2011. In the group of before 2010 of disease of cervical and lumbar spine disorder, the proportion was 10%, almost same with the proportion in the whole objects. But the proportion in the group of after 2011 increase dramatically to 15%. This might support the third reason for the crew-members spine disease. This might support the idea that with high acceleration, there is higher proportion of cervical and lumbar spine disorder happened in the crewmembers.
The second common disease was hyperlipidemia with proportion of 8%(CI 5-11%). Hongjie Shen and colleagues reported that hyperlipidemia was among the rst ve diseases in crew-members [15] . Jishun Yang et al supported this by presenting the results that hyperlipidemia was the second high proportion disease during investigation group [6] . The reason for this might related to the increase of food calories in crew-members' diets. Therefore, there is great need for aviation doctors focus on hyperlipidemia both on prevention and treatment.
There are other three kinds of trend we would like to discuss in the sub-analysis by investigation year. First, vegetative nerve functional disturbance decreased in the group of after 2011. This might related to diagnosis accuracy. Disease used to be diagnosis with vegetative nerve functional disturbance now was grouped accurately to other diseases. Second, fatty liver increased dramatically in group of after 2011 compared with group of before 2010. This might correlated to the increase of calories in the diet of crewmembers with improvement of Chinese economy. Third, arrhythmia also showed high proportion of increase compared with group of before 2010. The proportion was only 2% in group of before 2010. And it increased to 6% in the group of after 2011. This might related to the improvement of diagnostic capability of aviation doctors with the year passing by.
Our work demonstrated the most 15 common diseases for Chinese crew-members. With these results, aviation doctors could focus more on the frequently happened diseases. Also the proportion of other diseases was showed in the results. Our work make a clear vision for future research and aviation work.
The disease spectrum of crew-members is the classical way to get started for health management of crew-members. Our investigations also showed the variation of disease spectrum with different investigation year. This could guide research and improvement direction of the health service support in the future.

Limitations
Our results has limitation. Diseases happened for Chinese crew-members but not over 6 reports were not analyzed in our research. These diseases included depression, anxiety, knee joint disease, polyp of gallbladder, aviation disease, hyperuricemia and renal cysts and so on. These diseases also happened in inpatient crew-members and need further analyzed in near future.

Conclusion
Our paper presented all 15 common disease of Chinese crew-members. cervical and lumbar spine disorder was the rst frequently happened disease in inpatient crew-members. The proportion of cervical and lumbar spine disorder in crew-members varied according to investigation year. Other diseases were also showed in meta-analysis work of inpatient crew-members. This disease spectrum could offer proper guide for health management of Chinese crew-members. Abbreviations CI Con dence interval.
Due to technical limitations, table 1 to 3 is only available as a download in the Supplemental Files section. Figure 1 selection of studies for inclusion in meta