Characterization of workers covered by a risk insurance company in Colombia who suffered amputation

ABSTRACT | Background: Scarcity of information on the sociodemographic, occupational and clinical characteristics of workers who suffer upper limb amputation hinders planning and implementing actions to improve their living and working conditions. Objective: To investigate the sociodemographic, occupational and clinical characteristics of workers covered by a risk insurance company in Colombia who suffered upper limb amputations and investigate their association with risk categories. Methods: Exploratory, descriptive, retrospective and analytical study of data relative to workers covered by a risk insurance company in Colombia who suffered upper limb amputation in the period from 1982 and 2017. Following approval by School of Medicine, CES University, and a risk insurance company, we developed a checklist to collect data on the variables of interest. We performed descriptive and bivariate analysis with 95% of confidence and error of 5%. Results: The largest proportion of workers who suffered amputation were men aged 36 to 55 years old and residing in the Andean region of Colombia. Amputations mainly followed incidents during the performance of tasks, particularly in jobs in the secondary economic sector and classified as with type 3 risk. Conclusion: We were able to identify some characteristics associated with accidents, including age, risk class, economic sector, care received and disability duration. The results point to the need for appropriate interventions for the benefit of workers and risk insurance companies.


INTRODUCTION
Work accidents pose a serious problem 1 inasmuch as they are likely to lead to temporary or permanent disability and might result in amputations -especially of the upper limb -which frequency is estimated to increase in the near future in countries such as the United States [1][2][3][4] . Indeed, following death, accidents are one of the most serious work-related incidents due to their potential to cause incapacity not only to work, but also for activities of daily living. As a result, work accidents represent a true socioeconomic and psychological disaster for the affected workers and their families 1,3-7 . According to the Colombian Ministry of Health and Social Protection, organizations routinely have to deal with workers' disabilities, which have significant impacts on both employers and health care institutions. Amputation of limbs needed to accomplish tasks further increases costs to organizations, since the overall goal of treatment and rehabilitation is to achieve adequate work reintegration 8 .
Traumatic amputations are the second leading type of amputations among the overall population. However, there are some issues relative to the availability of data, since most concern amputations of the lower limbs. For this reason, gathering scientific evidence relative to risk factors, severity, working conditions, characteristics of accidents and clinical aspects of workers who suffer upper limb amputations is necessary 1,[3][4][5]8,9 . One of the main problems in this regard in Colombia derives from lack of information on the profile of this population of workers. Research has mostly focused on descriptions of treatments, while the data needed to achieve a better sociodemographic, occupational and clinical characterization have been rather neglected, more particularly in the case of workers covered by risk insurance companies (RIC) 5,8,10 .
As a function of the aforementioned considerations, gathering evidence likely to contribute to the design and implementation of prevention programs against work accidents and to improve rehabilitation, as well as to identify some of the -still scarcely investigated -characteristics of traumatic amputations is relevant 11 .

METHODS
In the present exploratory, descriptive, retrospective and analytical study we analyzed secondary data retrieved from records relative to workers covered by a Colombian RIC who suffered amputation in the period from 1982 to 2017. As per the Colombian Ministry of Health Resolution no. 008430 the present can be categorized as risk-free research and was approved by the research ethics committee of CES University on 6 June 2018, as well as by the involved RIC.
We considered records of workers covered by this RIC, aged above 18 and who suffered amputation. We excluded records with more than 20% of missing data, not available in electronic format or which did not indicate the involved limb.
We designed a checklist including: To control for bias we only considered variables with less than 20.0% of missing data and reviewed the checklist format, coherence and accuracy. We performed a previous pilot test with the checklist to assess application, time required for review and data collection, order, clarity and coherence of variables.

DATA ANALYSIS
We performed descriptive univariate analysis of the distribution of absolute and relative frequencies for qualitative variables. Quantitative variables were subjected to descriptive statistics (measures of central tendency, position and dispersion) and the normal distribution of data was assessed with the K-S test. On bivariate analysis, type of amputation (upper limb: yes/no) was defined as the dependent variable and all other aspects as independent variables. The χ 2 test was used to investigate association between two dichotomous or polytomous qualitative variables. Strength of association was expressed as prevalence ratio (PR) and corresponding 95% confidence interval (95%CI). The relationship between the dependent and quantitative independent variables, as e.g. age and work disability duration, was assessed by means of the Mann-Whitney U test. All the tests were performed with 95% confidence level and error of 5%. The data were processed using software Epidat version 3.1, Excel 2013 and Word 2013.

RESULTS
A total of 245 workers covered by the involved RIC suffered amputation in the period from 1982 to 2017. Most of these workers were male (94.0%; n=229) as shown in Figure 1. The highest prevalence of amputations corresponded to workers aged 36 to 55 years old (51.8%; n=126) followed by those aged 18 to 35 (33.5%; n=81) ( Figure 1). Most of the affected workers (74.0%) resided in the Andean region of Colombia ( Figure 1).
About 90.0% of the incidents leading to amputation occurred during the performance of job tasks. Most incidents involved entrapment, blows or being stepped on (80.4%, n=197) followed by exposure to/contact with electricity. Most cases corresponded to tasks classified as with risk type 3, i.e. medium risk (52.2%) and about one fourth (24.5%) as high risk (types 4 and 5). Almost half of the workers who suffered amputation (46.9%) worked in the secondary economic sector and 6.9% in the primary sector (Table 1).

CLINICAL ASPECTS OF WORKERS WHO SUFFERED AMPUTATION
About 82.4% of the workers (n=202) suffered traumatic amputation of the upper limb (shoulder, arm, forearm, wrist and/or hand). Amputations involving the wrist or hand/finger(s) were the most frequent (48.2%) followed by those at the level of the elbow or forearm (20.4%).
Most workers received in-hospital surgical treatment (69.8%) or in an intensive care unit (20.4%). The largest proportion of workers remained disabled 185 to 365 days (42.9%) ( Table 2).

SOCIODEMOGRAPHIC CHARACTERISTICS ASSOCIATED WITH UPPER LIMB AMPUTATION
As shown in Figure 2, the workers who suffered upper limb amputation were significantly younger than those who underwent other types of amputation (p=0.003). There was statistically significant relationship between higher prevalence of upper limb amputation and younger age (p=0.000). The prevalence of this type of amputation was 53.0% higher among the workers aged 18 to 35 compared to those aged 56 or older (PR=1.53; 95%CI 1.15; 2.03) and 46.0% higher among those aged 36 to 55 (PR=1.46; 95%CI 1.09; 1.94) ( Table 3).
We found statistically significant association (p=0.004) between residing in the Pacific region and lower prevalence of upper limb amputation. The prevalence of upper limb amputation was 28.0% lower (PR=0.72; 95%CI 0.55; 0.95) for the workers who lived in the Pacific region compared to other regions and 25% lower than that of workers who lived in the Andean region. We did not find any difference in prevalence as a function of sex (Table 3).

OCCUPATIONAL AND ACCIDENT CHARACTERISTICS ASSOCIATED WITH UPPER LIMB AMPUTATION
We found statistically significant association (p<0.05) between upper limb amputation and incident setting, economic sector and geographic region of residence (   (Table 4). Although statistically non-significant, the prevalence of upper limb amputation was 16.0% higher among the workers who performed type 3 risk tasks (Table 4).
Disability was shorter for the workers who suffered upper limb compared to lower limb amputation, 271 (interquartile range: 171, p<0.001) versus 453 days, as shown in Figure 3. Overall, disability lasted less than

SOCIODEMOGRAPHIC CHARACTERISTICS
The frequency of amputations was higher among younger workers (84.2% 18 to 55 years old) with more than 50% corresponding to workers aged 36 to 55. These findings agree with those reported in a study PR performed in Illinois, USA, in which 74.0% of amputations corresponded to workers aged 24 to 54. Barouti 14 found that two-thirds of amputations corresponded to workers under 40, as was also reported by Maryland and other authors 4, [14][15][16] .
As in other studies 1,4,7,[14][15][16][17] also in the present one amputations were most frequent among men, 94.0%, thus a rate similar to that reported by Camacho-Conchucos 1 , 98.2%, but higher than that in the aforementioned study performed in Illinois 15 , 88.8%. These findings suggest that men still perform high-risk jobs liable to lead to amputation. As possible reasons, female workers might not have been yet included in all production activities or are more careful in jobs with higher risk of accidents. Priya Varma et al. 4 observe that in addition to age, also sex (biological condition) has a considerable role in traumatic amputations in general, since in their study risk of amputation was twice higher for the men. However, we did not find significant association between sex and prevalence of upper limb amputations. Also differing from Priya Varma et al. 4 in our study the prevalence of amputation was highest (53.0%) for the youngest workers (18 to 35 years old) followed by the group aged 36 to 55 (46.0%).
The smallest proportion of workers who suffered upper limb amputation (28.0%) resided in the Pacific region (p<0.05). We were not able to locate any study that considered the place of residence of workers who underwent amputation.

OCCUPATIONAL CHARACTERISTICS
Most incidents leading to amputation took place during the performance of job tasks (88.5%) and only a smaller fraction while commuting. This rate is similar to that reported by Whelan et al. 18 , 85.5%. Incidents mostly involved entrapment, blows or being stepped on (80.4%) as also in the study by Camacho-Conchucos 1 in which the most common accidents involved entrapment (58.3%) followed by blows (12%).
Most incidents leading to upper limb amputation occurred within the secondary economic sector (46.9%) which precisely includes the occupations most exposed to entrapment, blows or being stepped on. In Camacho-Conchucos' study 1 the highest rate of amputations corresponded to the manufacturing sector (37.9%) therefore similar to our findings. However, he did not detail the activities included in this sector, therefore we cannot assert whether they were or not exactly the same as the ones we considered. Diverging from both Camacho-Conchucos 1 and our findings, in a study performed in Asturias, Spain 5 most workers who suffered amputation worked within the primary economic sector, versus 6.9% in our study.
We were not able to locate any study that analyzed variables geographical region and risk class for the purpose of comparison. The reason might derive from the particularities of the political-administrative division of Colombia and the risk classification used in the country, based on the Decree no. 1607 from 2002 12 . In our study 73.5% of the workers who suffered amputation lived in the Andean region (n=180), most (52.2%) had medium-risk jobs (class 3) and one fourth high-risk jobs (classes 4 and 5) as per the categories established in the just mentioned decree 12 .
We found statistically significant association (p=0.000) between upper limb amputation and economic sector. Prevalence was higher for workers in the secondary (39.0%) and primary (25.0%) sectors or had temporary jobs (31.0%) compared to the tertiary sector.

CLINICAL CHARACTERISTICS
Evidence in the literature indicates that most amputations involve the lower limbs 4,19,20 . However, in several studies more than 50.0% of amputations related to work accidents involved the upper limbs, 30.0% the lower limbs and 12% more than one limb 16,17 . In the present study, the prevalence of upper limb amputations was 30.0% higher, 82.4% (n=202) of them corresponding to traumatic amputations. This is a highly relevant finding, while this type of amputation might be related to risk categories and how job tasks are actually performed.
A considerable proportion of the workers who were victims of accidents (20.3%) required intensive care and 70.0% in-hospital surgical interventions. We found statistically significant association (p=0.000) between main health care approach and upper limb amputations. The latter's prevalence was higher among the workers who required outpatient (67.0%) or in-hospital (48.0%) surgical care. The prevalence of upper limb amputations was also associated (p=0.000) with disability duration. Prevalence was higher for the workers who remained disabled 91-180, 31-90 and 19-365 days. However, this type of information is still scarcely reported in the literature and this fact hindered our attempts at comparing results. In a study performed in Spain 5 70% of amputations were transradial and 30% transhumeral. In the study by Tennent et al. 19 transradial amputations corresponded to 47% of the total, the transhumeral level to 34%, wrist disarticulation to 13%, shoulder disarticulation to 4%, while elbow disarticulation was the least frequent (1%). In the study by Chul Ho Jang et al. 20 transradial amputations represented 51.9% of the cases, followed by transhumeral amputations (32.8%) and shoulder disarticulation (8.2%). Our findings therefore differ, since amputations at the level of the wrist or hand/ finger(s) (48.2%) or of the elbow or forearm (20.4%) were the most common.
Among the limitations of the present study, the source of information hindered analysis relative to possible variables, as well as the availability of data. We did not have access to variables representing all factors potentially associated with the analyzed type of amputations, in addition to the scarce published information on this subject and the period selected for the study, i.e. more than two decades.

CONCLUSION
Despite its limitations and difficulties in the development of the present study, we stress its significance as an onset of efforts to gather scientific evidence from the available information to thus facilitate later studies. More thorough studies are needed to establish the forms of and circumstances under which work accidents occur, with the known consequences for the lives and happiness of workers and their families, to thus ground actions from the perspective of workplace health promotion rather than centered on secondary and tertiary prevention, as was the case of the workers analyzed in the present study.