Background
Poverty is a widespread social phenomenon that not only exists in developing areas but also in developed regions [
1]. Although social welfare expenditures expand continuously with changes to the social environment, poverty still exists. Demands for governmental social assistance and related welfare benefits were significantly stronger in the low-income population (LIP) [
2]. The LIP is more subject to seriously disease than is the normal population (NP) in Taiwan [
2-
4]. Thus, it is necessary to conduct in-depth research and analyses to understand the disease condition and the cause, which can lead to suggestions for medical research institutions and governments. Appendectomy is one of the most common procedures worldwide [
5], and both rich and poor patients are likely to undergo it. Therefore, for this paper, we adopted appendicitis and appendectomy as the entry point to analyze the disease conditions in LIP patients.
Although numerous epidemiological studies on appendicitis have been conducted worldwide [
6-
13], only a few have focused on the effect of socioeconomic status (SES) on appendicitis and appendectomy [
14-
17]. Studies conducted specifically on the LIP are rare. Certain studies have also been conducted in Taiwan regarding the epidemiological features of appendicitis [
18-
25]. These studies concerned chiefly the monthly incidence variations of acute appendicitis [
19], the volume-outcome relation of acute appendicitis [
20], trend differentials in incidence rates for ruptured appendicitis between rural and urban populations [
21], and a comparison of the perforation rate of acute appendicitis between nationals and immigrants [
22]. However, epidemiological data on appendicitis specifically for the LIP have yet to be reported in Taiwan.
This study investigates the epidemiological features, particularly age and gender, length of hospital stay (LOS), hospital cost, incidences, and seasonal variations for the LIP in Taiwan. We compared these data with the epidemiological features of the NP to determine the effect of SES on appendicitis and appendectomy. We retrieved all of the data from the National Health Insurance Research Database (NHIRD) for 2003–2011.
Discussion
According to the provision of Section 13 and Paragraph 3 of the Taiwan Social Assistance Act, “The competent authority must hold living conditions survey for low-income population and moderate low income population at least every five years, and publish statistical reports” [
30]. The Taiwan Ministry of the Interior had conducted six surveys regarding the living conditions in the LIP and moderate LIP. The first survey was conducted in 1990, and the latest was performed in 2013 [
31]; the statistical reports from 2013 have not been presented yet. According to the survey results of 2004, the five main reasons for people becoming a part of the LIP were as follows: all family members are unable to work (35.33%), long-term illness (26.93%), many people in the family are unable to work (24.22%), other reasons (11.56%), and income earners who are divorced or separated (11.03%) [
2]. Therefore, long-term illness is the second leading cause of people becoming poor in Taiwan. In other words, long-term illness led to 62,554 people becoming a part of the LIP in 2004 (the LIP comprised 232,284 people in 2004). The survey results also showed that the proportion of families who had family members with a chronic or catastrophic illness reached 62.09% in low-income families in 2004, when the total number of families in the LIP was 78,428. Among them, families with one patient accounted for 51.73% of the total number of low-income families, families with two patients was accounted for 8.57%, and families with three or more patients accounted for 1.79%. The survey results in 2008 were similar to those of 2004 [
4]. For example, the proportion of families who had family members with a chronic or catastrophic illness reached 65.24% in low-income families in 2008, which was 3.15% higher than as indicated in the survey results of 2004. The situation in the LIP with disease was more serious than in the NP; hence, in-depth research and analyses is necessary for understanding the disease condition and the cause, to provide suggestions for medical research institutions and the government.
Previous studies have provided different definitions of appendicitis. For example, certain studies have defined a diagnosis of appendicitis as patients who had undergone an appendectomy [
5,
32,
33]. David et al. [
10] proposed that a patient with a positive primary appendectomy was considered to have acute appendicitis; the terms were used interchangeably in their paper. Lee et al. [
6] defined appendicitis as acute appendicitis (K35), other appendicitis (K36), and unspecified appendicitis (K37) according to the ICD-10. The definition in this study was similar to that by Lee et al. [
6], who applied a diagnosis of appendicitis, regardless of whether subjects underwent an appendectomy. This definition can more accurately distinguish between appendicitis, acute appendicitis, and appendectomy. However, it increases the incidence of appendicitis, causing it to be slightly higher than when the other aforementioned definitions are used. In the present study, we focused on comparing the LIP and the NP, and used the same definition for both groups; therefore, our results were not significantly affected.
The overall incidences of appendicitis, acute appendicitis, appendectomy, and perforated appendicitis in the LIP were 36.25%, 35.33%, 33.00%, and 37.28% higher than those in the NP, respectively. In other words, the risks of all aforementioned appendicitis in the LIP were higher than in the NP; therefore, an appendectomy was also more frequently performed in the LIP. To explore the reasons for the higher incidence in the LIP, we reviewed certain etiologic hypotheses on appendicitis to determine any association.
Numerous hypotheses have been proposed to explain the etiology of appendicitis, but only three of them have a measure of credibility and warrant further discussion [
34]. The first etiologic hypothesis was the mechanical hypothesis, which was proposed by Short [
35] in 1920. He hypothesized a causal relationship of appendicitis with a low cellulose content of imported foods. To the best of our knowledge, although the diet quality in the LIP may be less favorable than in the NP, differences in eating habits between the LIP and the NP were not substantial in Taiwan. Therefore, the possibility that the diet in the LIP contains more low-fiber foods compared with the NP was low. In addition, Barker and Liggins [
36] found that, despite similar dietary habits, the distribution of appendicitis did not follow other diseases associated with low fiber consumption. Therefore, there is a small possibility of a low-fiber diet leading to the higher incidence in the LIP than in the NP, and thus, we excluded eating habits as the reason for the higher incidence of appendicitis in the LIP than in the NP. The second etiologic hypothesis was the infection hypothesis; specific infections with viruses, bacteria, and parasites have been linked to appendicitis, prompting the suggestion that a local invasion could trigger appendicitis [
37-
40]. We agree with the infection hypothesis because it is possible that the LIP is more likely to be infected than the NP, resulting in the higher incidence of appendicitis in the LIP than in the NP. The third hypothesis is the hygiene hypothesis. The effect of better socioeconomic conditions because of improved water supplies and hygiene conditions has been found to be a reason for the decrease in the incidence of appendicitis [
41,
42]. We conjectured two possible reasons for the higher incidence of appendicitis in the LIP than in the NP, as follows: the LIP is infected more easily, and the LIP is under less favorable hygiene conditions. However, these conjectures warrant further research and in-depth clinical trials for verification, which we plan to conduct in the next phase of work. Our findings also revealed that the mean LOS for LIP patients with appendicitis, acute appendicitis, primary appendectomy, perforated appendicitis, and incidental appendectomy was 13.12%, 10.94%, 7.84%, 14.79%, and 7.18% longer compared with NP patients, respectively. This may be caused by three reasons. First, LIP patients may live in a more remote area than NP patients do; they tend to be uninsured, and may need to travel farther than NP patients do to obtain medical care [
21]. This may lead to a serious disease by the time they arrive at a hospital because of the delay, and hence, they may need a long LOS. This may also be the reason that a higher incidence of perforated appendicitis was found in the LIP compared with the NP. Second, poor financial conditions may result in a poor quality of life, and therefore, the constitution of the LIP may be weaker than that of the NP, thereby requiring a lengthier recovery time after an appendectomy. Finally, a certain relationship with the health care system in Taiwan may affect the incidence as well. Because LIP patients are not required to pay any fees covering hospital costs according to health care provisions in Taiwan, certain LIP patients may be less likely to consider payment problems for long LOS when they use medical resources.
Based on our experience, most appendectomy was caused by acute appendicitis, which is verified in our study (94.27% for LIP patients, 93.50% for NP patients, respectively). However, we also found that some patients who underwent appendectomy were diagnosed with ICD codes of 541–543 (unqualified appendicitis, other appendicitis or other disease of the appendix) (2.79% for LIP patients, 2.87% for NP patients, respectively). Because the surgery for chronic appendicitis is rare, this situation is an interesting phenomenon. The primary reason for this is that some patients may require appendectomy even when the symptoms for acute appendicitis are not obvious. Some physicians’ improper coding behavior may also lead to this situation, which needs further study to clarify.
LA is not routinely performed for appendicitis because the operation costs associated with that procedure are higher than those associated with OA. However, our findings revealed that the total cost was comparable between the LA and OA (1,191 ± 19 USD of OA versus 1,178 ± 13 USD of LA,
p < 0.05) in LIP patients. The mean LOS of OA for LIP patients was longer than that for NP patients (5.51 ± 0.11 versus 5.17 ± 0.01,
p < 0.05), but it was similar for LA between LIP and NP patients (3.80 ± 0.08 versus 3.80 ± 0.01,
p < 0.05); hence, more hospitalization costs are saved when LIP patients chose the operation type of LA. In terms of hospital costs and LOS, LIP patients benefit more from the LA approach for the treatment of appendicitis. Nevertheless, more prospective investigations should be designed to explore the economic advantages of LA, such as the time back to work and normal daily activity [
23].
The overall incidence of appendicitis, acute appendicitis, and primary appendectomy appeared to be strongly age related in both the LIP and the NP, with the highest incidence in those aged 15–29 years, but a lower incidence in the younger and older age groups. In addition, the incidence of perforated appendicitis appeared to be age related in both the LIP and the NP, and was highest in older people and lower in younger people. The perforation ratio was also strongly age related in both population groups, and was highest in older people and lowest for the age group of 15–29 years. This phenomenon has also been observed in previous studies [
6,
10,
43,
44], in which the researchers had called it “J-shaped”. Furthermore, some of these studies have divided the 0-14-years old group into three groups; they found that the perforation ratio of 0–4 years was extremely high. By referring to their classification method that divided the age group into three age groups, our data also revealed the same characteristics (the data are not presented in the paper). As David et al. [
10] stated, this pattern reflects both the increased diagnostic difficulty and less timely surgical intervention for people in these extreme age groups.
Regarding seasonal variations, although the incidence of appendicitis in the LIP did not exhibit an obvious regularity as it did in the NP, we observed that the incidence was slightly higher in summer than in winter based on the overall data. This pattern has been observed in previous studies as well [
6,
10,
19,
33,
45]. Wei et al. [
19] analyzed the relationship between the incidence of appendicitis and climate factors, including ambient temperature, relative humidity, atmospheric pressure, rainfall, and hours of sunshine, and they found that only the ambient temperature was positively correlated with the incidence of appendicitis. Kaplan et al. [
46] reported a significant effect of air pollution on the incidence of appendicitis in the summer months. Several factors may contribute to the seasonality of appendicitis and appendectomy, but no single causative factor has been identified [
6,
10,
32].
The overall incidence of incidental appendectomy in the LIP was 8.69 per 100,000 per year, which was 11.4% higher than in the NP (7.80 per 100,000 per year); however, this is lower than what has been found in certain studies [
10]. The overall incidence of incidental appendectomy decreased by 43.3% between 2003 and 2011, and the decline trend is consistent with the findings in previous studies [
10], but the decline ratio was greater.
The NHIB has established a uniform system to control the quality of medical services and coding, and therefore, the quality of data acquisition in the present study was reliable [
20,
47,
48]. However, our data are still subject to limitations. In total, 1,240 records of appendicitis patients were missing information regarding sex (six LIP patient records and 1,234 NP patient records) between the years 2003 and 2004; in one record the sex information was absent in 2006 and 2010. However, sex information for the other years was complete. The missing sex information did not affect the calculation of the overall incidence unrelated to sex information, but certain errors are to emerge when we conduct a comparison of the incidence in males and females at different ages. To resolve this problem, we calculated the number of records for male and female patients in each age group because those records contained sex information, and then we divided the number of males by that of females to obtain a male–female ratio. Afterward, we randomly assigned records of the same age group without sex information to the male or female groups according to the obtained sex ratio. This solution retained the total number of records as unchanged, and ensured that the male–female ratio was relatively accurate, but it still resulted in some deviation, which is one drawback of our study.
Competing interests
All authors declare that they have no conflicts of interest, including directorships, stock holding or contracts.
Authors’ contributions
The study was designed by KBL and NPY; data were gathered and analyzed by RKL and RHP; the initial draft of the manuscript was written by CLC, KBL, and SZZ; the accuracy of the data and analyses was assured by YHL and KBL. All authors participated in the preparation of the manuscript and approved the final version. All authors read and approved the final manuscript. In general, KBL, CLC, NPY and RKL contributed equally to this work.