Shoulder Dislocation Incidence and Risk Factors—Rural vs. Urban Populations of Poland

(1) Background: The aim of this study was to analyze the incidence of shoulder dislocation and to estimate non-modifiable risk factors in rural and urban subgroups in Poland. (2) Methods: The study covered the entire Polish population, divided into urban and rural subgroups and observed between 1 January 2014 and 31 December 2014. The study population consisted of Polish patients with a diagnosis of shoulder dislocation (S43.0) in accordance with the International Statistical Classification of Diseases and Related Health Problems (ICD-10). Records were obtained from the public health care provider National Health Found (Narodowy Fundusz Zdrowia, NFZ). Based on these data, we assessed shoulder dislocation incidence and risk rates, stratifying the study sample by sex, age and place of residence (rural or urban) using the Central Statistical Office (GUS) personal territorial code (TERYT). (3) The incidence was 25.97/100,000 person-years in rural areas and 25.62/100,000 person-years in urban areas. We did not find significant differences in the incidence between the two subgroups. The highest incidence (75.12/100,000 person-years) and the highest risk for shoulder dislocation were found among subjects 80+ years old living in urban areas. Furthermore, men in the third decade of their life living in urban areas showed the highest risk (OR = 7.8, 95% CI; 6.44–9.45, p < 0.001). In both subgroups, the likelihood of shoulder dislocation was significantly lower for the female sex and among children ≤9 years old. However, girls living in rural areas presented with a significantly higher likelihood for dislocation compared with their peers living in urban environments. (4) Conclusions: No significant difference in the incidence rate of shoulder dislocation between Polish residents living in rural and urban areas emerged. The highest incidence was observed among female subjects 80+ years old living in urban environments. The highest risk was found among men in the third decade of their life living in urban areas. In addition, girls in the first decade of their life living in rural areas had more shoulder dislocations than girls living in urban environments. Shoulder dislocation is dominant in female subjects aged 70–79 living in rural areas and in females 80+ years old living in urban areas.


Introduction
Shoulder dislocation is a trauma that nearly all emergency department doctors confront at some point during their professional careers due to its high incidence. In Poland, Int. J. Environ. Res. Public Health 2022, 19, 11857 2 of 15 the incidence of shoulder dislocation is 26.69/100,000 person-years [1]. Based on the data available in the literature, incidence rates range from 8.2 to 26.2/100,000 person-years [2,3]. Despite discrepancies in the data, there is consensus that the shoulder joint is the most frequently dislocated joint in the human skeletal system [2,[4][5][6]. Considering modifiable risk factors, athletes are one of the most vulnerable subgroups, in particular athletes practicing contact sports and sports that require the arms moving to an overhead position [7,8].
Of all joints in the human body, the shoulder joint is the most mobile, due to its unique anatomical structure: the articular surface of the humeral head is indeed three times larger than the articular surface of the acetabulum, allowing for a wider range of movement [9]. Joint stabilizers (both static and dynamic) further impact the joint's mobility, affecting the entire upper limb's ability to grasp and manipulate objects with adequate efficiency [10]. On the other side, this high mobility is also responsible for the increased predisposition to dislocations.
Epidemiological studies are missing on shoulder dislocations based on metadata analyses describing or comparing entire populations or their significant subgroups beyond risk groups. Our aim was to start addressing this gap by comparing the incidence of shoulder dislocation in rural and urban subgroups in Poland, which also allowed for an indirect assessment of the existing non-modifiable risk factors in both subgroups as well as a first evaluation of the correlation between place of residence and health care availability.

Identification of the Test Groups
The National Health Fund (NFZ) keeps a register of all medical diagnoses coded in accordance with the International Statistical Classification of Diseases and Related Health Problems (ICD-10) throughout Poland. We screened the data collected by this fund between 1 January 2014 and 31 December 2014 to identify patients with posttraumatic dislocation of the shoulder using ICD-10 code S43.0 (dislocation of the shoulder joint). Only four-digit codes precisely related to the specified diagnosis were considered. Each patient with a dislocation was given a unique numerical identifier, for reasons of anonymity and for the protection of personal data.

Population Data
As the reference group, we considered the entire population of Poland not suffering shoulder dislocation in 2014 (from 1 January 2014 to 31 December 2014). Basic population data such as number of inhabitants per town and village using territorial code (TERYT), age and sex were obtained from the Central Statistical Office (GUS), the national agency for collecting population data, including demographics, in Poland [11].

Inclusion and Exclusion Criteria for the Test Group (Patients with Shoulder Dislocation) 2.3.1. Inclusion Criteria -
Polish residents as verified by their territorial code; -One-off dislocations of the shoulder coded during the studied period; -Diagnoses performed in emergency rooms and coded according to the ICD-10 classification.

Exclusion Criteria
-Shoulder dislocation diagnoses not coded in accordance with ICD-10; -Consecutive shoulder dislocations for the same patient, considered recurrences; -Patients not residing in Poland.
Based on the above criteria, we obtained a group of patients who reported to an emergency department with a single posttraumatic shoulder dislocation diagnosis during the period under consideration.

Determination of the Incidence and Risk Rates for Shoulder Dislocation
Using the NFZ data, in which each S43.0 diagnosis was accompanied by a territorial code, the study group was divided into two subgroups: rural and urban. Patients were further stratified based on age and sex. The demographic data were analyzed, and for each of the outlined groups, as well as for the entire study population, the number and frequency of posttraumatic shoulder dislocations over the studied period were determined. The incidence rate was expressed as the number of new registered cases per 100,000 personyears. These results were used to compare the relative incidence rates in rural and urban subgroups both as a whole and stratified for age and sex. To estimate the risk of shoulder dislocation, we calculated the odds ratios (ORs) of urban vs. rural shoulder dislocation for the entire population and for the age and sex subgroups as well as the risk of dislocation in rural vs. urban environments.

Statistical Analysis
Microsoft Excel software v2002 (compilation 1252.22215) (Microsoft Corporation, Redmond, WA, USA) was used to store data and generate charts. Differences in the frequency of shoulder dislocation in regard to gender and age ranges were assessed with Pearson's chisquare (χ 2 ) test. Odds ratios (OR) with 95% confidence intervals (CI) were calculated using Epi Info™ 7.2 software, developed by the Centers for Disease Control and Prevention (Atlanta, GA, USA). Statistica 12.0 software (STATSOFT, Tulsa, OK, USA) was used in all of the above analyses, and p-values less than 0.05 were considered to be statistically significant.

Results
The population of Poland for the observed period amounted to 38,478,602 individuals, among whom, 15,262,250 (39.67%) resided in rural areas, with a female population of 7,652,607 (50.14%) and a male population of 7,609,643 (49.86%), while the remaining 23,216,352 (60.33%) people resided in urban areas, with a female population of 12,206,186 (52.58%) and a male population of 11,010,166 (47.42%).

Analysis of the Number of Shoulder Dislocations
During the study period, 9912 shoulder dislocations were registered: 3964 (39.99%) dislocations took place in rural areas, of which 2602 (65.64%) occurred in men and 1362 (34.36%) in women. The other 5948 (60.01%) dislocations were reported in urban areas, including 3745 (62.96%) dislocations occurring in men and 2203 (37.04%) in women. The highest number of dislocations was reported in patients 60 to 69 years old both in rural and urban subgroups, with 746 and 1100 dislocations respectively. The details are presented in Figure 1.
Stratifying the groups by age and sex, the largest total number of incidents was found in men in their third decade of life, both in the countryside and in the city, with 505 and 782 dislocations, respectively ( Figure 2).
For the female population, the highest prevalence of dislocations was found in the 60-69 age subgroup living in rural areas (24.89%). For the male population, the 20-29 age subgroup residing in urban areas presented the highest prevalence (20.88%) of dislocations. The dislocation prevalences for all age subgroups analyzed are reported in Figure 3.

Incidence Rates in Urban and Rural Subgroups
The incidence rate was 25.97/100,000 person-years in rural areas and 25.62/100,000 person-years in urban areas. The highest incidence rate was found in people older than 80 (64.64/100,000 person-years) living in urban environments, while in rural areas, the 70-79 age group presented the highest rate of dislocations (62.71/100,000 person-years). A comparison of the incidence rates in rural and urban subgroups is shown in Figure 4.    Stratifying the groups by age and sex, the largest total number of incidents was found in men in their third decade of life, both in the countryside and in the city, with 505 and 782 dislocations, respectively ( Figure 2).   Taking into account age and sex, the highest incidence rate was found in female city residents 80+ years old (75.12/100,000 person-years). Among men, the highest incidence rate was found in the 60-69 age group living in rural areas (62.59/100,000 person-years). Incidence rates by sex are presented in Figure 5.
For the female population, the highest prevalence of dislocations was found in the 60-69 age subgroup living in rural areas (24.89%). For the male population, the 20-29 age subgroup residing in urban areas presented the highest prevalence (20.88%) of dislocations. The dislocation prevalences for all age subgroups analyzed are reported in Figure  3.

Incidence Rates in Urban and Rural Subgroups
The incidence rate was 25.97/100,000 person-years in rural areas and 25.62/100,000 person-years in urban areas. The highest incidence rate was found in people older than 80 (64.64/100,000 person-years) living in urban environments, while in rural areas, the 70-79 age group presented the highest rate of dislocations (62.71/100,000 person-years). A comparison of the incidence rates in rural and urban subgroups is shown in Figure 4.

Incidence Rates in Urban and Rural Subgroups
The incidence rate was 25.97/100,000 person-years in rural areas and 25.62/100,000 person-years in urban areas. The highest incidence rate was found in people older than 80 (64.64/100,000 person-years) living in urban environments, while in rural areas, the 70-79 age group presented the highest rate of dislocations (62.71/100,000 person-years). A comparison of the incidence rates in rural and urban subgroups is shown in Figure 4.  No significant correlation (p = 0.505) between the number of dislocations and the place of residence emerged from the analysis. However, the number of dislocations correlated significantly with age (p < 0.001) and sex (p < 0.001). Taking into account age and sex, the highest incidence rate was found in female city residents 80+ years old (75.12/100,000 person-years). Among men, the highest incidence rate was found in the 60-69 age group living in rural areas (62.59/100,000 person-years). Incidence rates by sex are presented in Figure 5. No significant correlation (p = 0.505) between the number of dislocations and the place of residence emerged from the analysis. However, the number of dislocations correlated significantly with age (p < 0.001) and sex (p < 0.001).

Risk Assessment of Shoulder Dislocation Depending on the Place of Residence
The analysis of the odds ratios for the rural and urban subgroups independently (Tables 1 and 2) showed that the lowest expected risk of dislocation was in the 0-9 age subgroup of the female population, especially among girls living in urban environments. The reference group consisted of the whole female population (female residents: OR = 0.02, 95% CI; 0.01-0.05 p < 0.001). The risk of dislocation started to increase in the sixth (rural population) and seventh (urban population) decades of life and continued to increase in the oldest age groups. For both subgroups, the highest risk of dislocation was among the two oldest age groups. The highest risk of dislocation was recorded in women in the 80+ age group in urban areas and the 70-79 age group in rural areas. In the rural population, the highest risk for men in the rural population was found in the age ranges 60-79 years, and it was in the age ranges of 20-29 and 70-79 for men living in urban areas. Odds ratios for both subgroups are detailed below (Table 1, rural population; Table 2, urban population).

Risk Assessment of Shoulder Dislocation Depending on the Place of Residence
The analysis of the odds ratios for the rural and urban subgroups independently (Tables 1 and 2) showed that the lowest expected risk of dislocation was in the 0-9 age subgroup of the female population, especially among girls living in urban environments. The reference group consisted of the whole female population (female residents: OR = 0.02, 95% CI; 0.01-0.05, p < 0.001). The risk of dislocation started to increase in the sixth (rural population) and seventh (urban population) decades of life and continued to increase in the oldest age groups. For both subgroups, the highest risk of dislocation was among the two oldest age groups. The highest risk of dislocation was recorded in women in the 80+ age group in urban areas and the 70-79 age group in rural areas. In the rural population, the highest risk for men in the rural population was found in the age ranges 60-79 years, and it was in the age ranges of 20-29 and 70-79 for men living in urban areas. Odds ratios for both subgroups are detailed below (Table 1, rural population; Table 2, urban population).

Dislocation Risk in Rural vs. Urban Subgroups
The comparison of the risk (OR) for the first event of shoulder dislocation among urban and rural subgroups, taking into account the stratification by age and sex, unraveled that girls aged 0-9 years living in the countryside are the most exposed to shoulder dislocation accidents. Furthermore, the risk of dislocation was shown to progressively increase in both sexes between the ages of 50 and 80 in rural and urban subgroups, with the exception of women in their sixth decade of life. The comparison of the age-related odds ratios (ORs) for the first shoulder dislocation event within rural and urban subgroups is presented in Table 3.

Risk Assessment of Shoulder Dislocation Depending on Place of Residence and Sex
There is an increased risk for the occurrence of the first shoulder dislocation events in men of both subgroups (urban areas: OR = 1.7, 95% CI; 1.61-1.79, p < 0.001; rural areas: OR = 1.92, 95% CI; 1.8-2.05, p < 0.001). The comparison of the likelihood for dislocation taking into account sex and age group showed that men in their 20s living in urban areas share the highest risk (OR = 7.8, 95% CI; 6.44-9.45, p < 0.001). Moreover, a significantly high risk was found also in the subgroup of men aged 30-60 years old living in rural areas (see Table 4).

Discussion
In this study, the incidence rates of shoulder joint dislocation were 25.97/100,000 personyears in the rural subgroup, and 25.62/100,000 person-years in the urban subgroup. The highest incidence rate (75.12/100,000 person-years) was found among women aged 80+ years old living in urban areas, while women 60-69 years old living in rural areas shared the highest prevalence of shoulder dislocations within the female population. On the other hand, while the lowest event rate was found among girls aged 0-9 years old for both rural and urban subgroups, a significantly higher risk for dislocation associated with living in rural areas emerged from the differential comparison based on place of residence.
The main results of this study are in line with those obtained previously from the analysis of metadata of the Polish population that did not take into consideration the place of residence (26.69/ 100,000 person-years) [1], as well as with data presented by other authors, e.g., Zacchilli (USA) (23.9/100,000 person-years) and Leroux (Canada) 23.1/100,000 person-years [3,12]. Both authors did not investigate the dislocation incidence in correlation with the patients' age or the risk extent.
No statistically significant differences emerged from the analysis of the incidence rates for urban and rural subgroups. The incidence rates in the male and female populations living in rural areas were 34.19/100,000 person-years and 17.8/100,000 personyears respectively. Similar results were recorded for both sexes in the urban population (34.01/100,000 person-years and 18.05/100,000 person-years, respectively, for male and female subgroups). The incidence rates for both genders was similar to those reported in epidemiological studies from other authors: Liaavag et al., reported incidence rates of 34.8/100,000 person-years in the male population and 17.9/100,000 person-years in the female population [13], while Zacchilli et al., found incidence rates of 34.9/100,000 personyears for men and 13.3/100,000 person-years for women [3].
Significantly higher incidence rates were found in studies of risk groups and/or performed on small populations. For soldiers, Amako et al., reported an incidence rate of 410/100,000 person-years [14], and Kardouni et al., a rate of 313/100,000 person-years [15]. In another study, the frequency of shoulder dislocation in players practicing rugby was 100/100,000 person-years in men, and 50/100,000 person-years in women [16]. Recent publications (2020-2022) presented data from studies of rugby players, American football players, and a group of children and teenagers [17][18][19][20][21]. These incidence rates are peculiar to these specific subgroups and not representative of the general population. Studies conducted on small risk groups bring extremely important and useful information about these subgroups of people, though the results cannot be extrapolated to the general population, and their predictive value for other/larger groups is limited.
Unfortunately, there is a lack of epidemiological studies on wider, heterogeneous groups, e.g., large populations living in urban and rural environments, and studies comparing data from such subgroups. In 1984, Simonet et al., for the first time analyzed the incidence of shoulder dislocation in urban and rural populations using place of residence as a criterion for comparison. The study object was the population of Olmsted, Minnesota. The incidence in the city amounted to 7.6/100,000 person-years in urban areas and 10.2/100,000 person-years in rural areas [22]. We attribute the differences between these numbers and ours to the different study group sizes. Simonet analyzed 116 patients with shoulder dislocation, while our results are derived from the analysis of 9912 dislocations overall. Moreover, the study By Simonet et al., was performed in the 1970s on a population of 880,000 residents, while our subjects were the entire Polish population (38,478,602 people) [22]. Despite these differences, Simonet et al., as in our case, did not find statistically significant differences between incidence rates in rural and urban areas [22]. Shields et al., show relatively up-to-date data on the urban population of the UK, for which an incidence rate of 21.9/100,000 person-years is reported [23]. Shields et al., report a bimodal incidence peak for men aged 15-24 (421/100,000 person-years) and 85+ (509/100,000 person-years) and a peak incidence of 457/100,000 person-years in men aged 65-74 [23]. In the present study, a similar bimodal/unimodal incidence peaks trend was observed. However, peak incidences were found in different age ranges for our Polish urban subgroups. Peak incidence rates for men were found at ages 20-29 and 70-79 (49.69/100,000 person-years and 52.79/100,000 person-years respectively), while for women, the peak was unimodal with 75.12/100,000 person-years at 80+ years of age. In our opinion, the reasons for these differences lay in the inclusion criteria for the study-our study examines the entire population, while Shields only includes inhabitants over 15 years of age. Moreover, the differences presented above may be influenced by the higher life expectancy of men in the UK [23].
An analysis of groups without age and sex stratification showed that the place of residence (urban vs. rural) was not a risk factor for the first event of shoulder dislocation (OR = 1.01, 95% CI; 0.97-1.05; p = 0.043; p = 0.504). With our analysis of the risk of shoulder dislocation based on sex and pre-defined age groups, we could identify differences in the risk distribution among groups. The highest risk was found for men between 20 and 29 years old living in urban areas (OR = 7.8, 95% CI; 6.44-9.45, p < 0.001). For the age range 30-60 years, the highest risk is in men living in rural areas. This is in agreement with other data available from the medical literature. We think that the higher risk of dislocation in this subgroup is due to the significant exposure to trauma associated with manual labor on a farm, but confirming this assumption would require a detailed study of a smaller population. Other authors also reported that in the urban population, the increased risk of dislocation among young men is a consequence of their frequent participation in sports [3,22,[24][25][26]. However, we want to emphasize that due to lack of comparable studies, the conclusions above require further verification. Similar results were obtained in the male (OR = 1.01, 95% CI; 0.96-1.06, p = 0.836) and female (OR = 0.99, 95% CI; 0.93-1.06, p = 0.685) subpopulations.
In summary, the highest risk for shoulder dislocation in the rural subpopulation was found in the 70-79 age group (OR = 4.17, 95% CI; 4.17-5.33, p < 0.001), while in the urban population, it was shared by women aged 80+ (5.07, 95% CI; 4.58-5.61, p < 0.001). While the propensity to injuries in older patients can be attributed to many factors (degenerative changes, balance disorders, decreased physical fitness), the high exposure we found in girls younger than 9 years old should prompt us to consider the introduction of both prophylactic measures against injuries at an early age as well as educational measures to promote awareness in adult caretakers. However, it should be emphasized that in both urban and rural subpopulations, belonging to the female gender and being in the age range 0-9 years old are both protective factors even though living in rural areas implies a 20-fold higher risk of dislocation than living in urban areas, which should raise some concern. According to the authors of this paper, this situation might result from the specific activities and lifestyle in rural areas: the living and working place may entail more risks, which both children and their caretakers may fail to recognize. According to epidemiological data, the incidence of injuries in children aged 0-9 living in the countryside is indeed strictly related to living and working in this specific environment (animals, agricultural machines).
The risk groups identified here may differ in terms of traumatic complications, diagnosis and treatment. Therefore, doctors working at emergency rooms should have basic skills consistent with the available knowledge on the epidemiology and management of patient with a dislocated shoulder in different age groups.
Our study is the first countrywide study as we analyze data of the entire Polish population, without exclusions based on, e.g., age or gender. The heterogeneity of such a large group averages out the real incidence of shoulder dislocation accidents in found in specific risk groups (like rugby players) and provides a good approximation of the risk for the general population. However, only non-modifiable factors like sex and gender were available in the NFZ database used for this study. The inability to identify the causes of dislocation and the patients' medical history (additional diseases, dislocations, past injuries, consequences of dislocations, applied treatment BMI, level of sports activity or type of work, etc.) accounts for some limitations and reduces the predictive power of these results for other populations/subgroups that might benefit from a specific, smaller group analysis instead.
In the future, it will be important to refine the analysis and provide more precise data as most emergency doctors will encounter shoulder dislocations. We believe that further population studies are necessary for gaining knowledge about the impacts of both non-modifiable risk factors and modifiable risk factors (like level of sports activity or type of work performed) on dislocations. Further studies should also provide information on the risk of recurrence and development of shoulder instability for all subgroups studied, which will complement the current knowledge in this field. We also believe that research based on smaller populations of towns and villages is necessary and will allow for a more fine-tuned representation of the incidence and risk distribution. The subgroup of girls aged 0-9 years living in rural areas also calls for more specific analyses to explain the significantly higher risk of dislocation observed here.

Conclusions
Based on our data, no differences in the incidence rates of dislocations in urban and rural subgroups can be identified, with the exception of girls in their first life decade.
On the other hand, the most important risk factor is represented by age and gender. In females, the highest incidence rate is found in the 70-79 age group in rural areas and in the 80+ age group in urban areas. Female gender and age in the first life decade are strong protective factors within both studied subpopulations, but girls aged 0-9 living in rural areas have a significantly higher risk of shoulder dislocation than those living in urban areas. Considering the entire Polish population, men in their third life decade have an almost eight times higher risk of shoulder dislocation than the rest of the population taken together.  Institutional Review Board Statement: Ethical review and approval were waived for this study due to retrospective observational character of study and containing irreversibly anonymized data.
Informed Consent Statement: Informed consent was waived for this study due to irreversibly anonymized metadata. Data Availability Statement: Data are available upon special request.

Conflicts of Interest:
The authors declare no conflict of interest.