Prevalence and factors associated with shoulder pain in the general population: a cross-sectional study

| Musculoskeletal pain in the shoulder is considered one of the most prevalent causes of pain and disability in adults. To verify the prevalence of shoulder pain and correlated factors in adults aged 20 years or older in cross-sectional study with 600 individuals interviewed through questionnaires: (1) participants characterization; (2) level of physical activity; (3) reported morbidities; and (4) musculoskeletal symptoms. Descriptive, bivariate, and Poisson regression analyses were performed. The prevalence of pain in the shoulder was 24% (CI 20.3%-27.5%). Being 60 years or older (PR=2.14; 1.33-2.45), female (PR=1.92; 1.29-285), using the computer more than three times a week (PR=1.55; 1.01-2.32), working in a sitting position (PR=1.64; 1.03-2.59), standing up leaning the body forward (PR=1.54; 1.00-2.37), and reporting two or more morbidities (PR=3.31; 1.97-5.57) were all indicators of shoulder pain. This study discloses a high prevalence of shoulder pain and a strong relation with women, age, those who use the computer more than three times a week, those who execute occupational activities sitting and standing up leaning the body forward, and those who report two or more diseases.

1.29-285), using the computer more than three times a week

INTRODUCTION
Shoulder pain is the third most common musculoskeletal condition in the primary health care system -with an annual prevalence between 4.7 and 46.7 1-3 -following the low back and neck pain, respectively, with 13% to 65% 4 and 16.7% to 75.1% 5 .
Studies show the relation between shoulder pain and some correlated factors, such as being women, performing manual movements above the shoulder level, using vibrating tools, sitting in a position with cervical spine flexion, work standing, performing hard and repetitive physical work, and using the computer daily [1][2][3][4][5][6][7] . Besides these biomechanical work constraints, psychosocial risk factors (high demand, lack of autonomy, and low social support) and psychological illnesses (depression and anxiety) have been identified [8][9][10] .
Shoulder pain is a musculoskeletal condition i.e. a significant cause of morbidity and functional disability in occupational and daily life activities, contributing to increased social and economic costs for individuals, businesses and the State 11 .
In Brazil and worldwide, there is little information available about the epidemiology and clinical characteristics of shoulder pain, mainly in population-based studies. Furthermore, current research about shoulder pain is predominantly related to work factors, not focusing on individual factors and life habits 12 .
The knowledge of the prevalence of shoulder pain, its clinical aspects, and risk factors are important, since musculoskeletal conditions affect populations of different age groups and sex. Thus, this knowledge will enable health professionals to provide patients with adequate information about the most likely course of their symptoms, prognostic information to distinguish between patients with favorable outcomes and those with a high risk of chronic shoulder pain and disability, and to implement curative and preventive interventions 13 .
This study aims to verify the prevalence of shoulder pain in a sample population of adults, aged 20 years or older, as well as analyze the correlations between shoulder pain and demographic, socioeconomic, and ergonomic aspects related to lifestyle and reported morbidities.

METHODOLOGY
This cross-sectional study was conducted in the urban area of Bauru. The age and sex groups, denominated sample domains -20 to 35 years-old men; 20 to 35 yearsold women; 36 to 59 ,years-old men; 36 to 59 years-old women; 60 year-old and older men; and 60 year-old and older women -were first defined with a minimum number of individuals per sample.
To calculate the sample size, was used a estimated proportion of 50% of the population subgroups with confidence level of 95% in the estimation intervals as well as a sampling error of 10%, and a design effect (DEFF) of 2%. Therefore, the sample size for each group was at least 200 individuals (100 men and 100 women), totaling 600 participants.
The sample was taken from a two-stage cluster -the primary sampling units (PSU) were the census sectors, and the secondary sampling units, the residences. The PSU were collected by systematic sampling with probability proportional to their size. The sampling units were obtained from the 2011 National Household Sample Survey (PNAD). In total, 50 urban census sectors were taken from the 476 identified sectors.
It was decided that about 12 families should be visited for each census sector by calculating the ratio of the average number of individuals to the number of households. Households were systematically drawn and all individuals residing in them were eligible. A new home was randomly selected in case of refusal. Individuals not found after four visits (of which at least one occurred at night and one on the weekend) were considered as loss. Those who refused to answer the questionnaire were considered as refusals.
Individuals who were living in institutions such as nursing homes and prisons and those who were unable to answer the questionnaire were excluded from the study. Some older adults underwent the Mini Mental State Examination and those who scored below 27 points were excluded from the study 14 . Data were collected from February to June 2012. Interviewers and senior Physical therapy students were submitted to theoretical and practical training.
The variable "shoulder pain" was assessed using the Nordic questionnaire, validated and adapted to the Brazilian culture 15 . Shoulder pain was defined as a pain located at a restricted area in or around the shoulder complex 16 . In the interview, individuals were asked: "Have you had any pain or discomfort in the shoulder area in the past year?" 15 .
Individual factors included sex, age, marital status, education, race and, income. The questions regarding devices electronics -TV, computer or playing video games -were "Do you watch TV during the week?"; "How many times do you watch TV during the week?"; "How many hours do you watch TV in a typical day?"; "How often do you use the computer or play video games in a typical week?"; "For how many hours do you use a computer or play video games on a typical day?" 17 .
The ergonomic variables -repetitive movements, heavy lifting, vibration/tremor during the working day, sitting position, forward leaning position, weighted sitting position, postural orthostatic, forward flexed, kneeling and lying posture -were characterized by the options that best describe their exposure frequency during a workday as follows: never, rarely, usually, always.
Smokers were those who reported smoking daily (at least one cigarette per day) or occasionally (less than one cigarette per day) or former smokers who had stopped smoking for at least six months prior to the interview 18 . Morbidities and health problems were assessed by selfreport of the presence in the last 12 months.
The level of physical activity was verified by the International Physical Activity Questionnaire (IPAQ) 19 , short version. For the classification of the subjects in the levels of physical activity, IPAQ official categories were used -insufficiently active, sufficiently active and very active 19 .
The data obtained were inserted into a database using SPSS statistical software, version 10.0 (SPSS, Chicago, United States). Absolute and relative frequency distributions were performed for categorical variables, as well as the calculation of prevalence ratios (PR) with 95% confidence interval (CI). Poisson regression was utilized to analyze the variables associated with shoulder pain, respecting a hierarchical model of relationships among the variables. The prevalence ratios were calculated, as well as their respective 95% confidence intervals.

RESULTS
The drawn residences included 641 eligible subjects, of which only 600 were effectively interviewed. The main reasons for loss (No.=41) were: "absent residents" and "scheduled with the interviewer but did not attend". Refusals were: "I do not respond to interviews" and "too long, it will take a long time to respond". Most participants had between 9 and 11 years of formal education, were white, married, of low income, non-smoker, and sedentary.
Regarding the total of individuals, 24% (CI 20.3% to 27.5%) of 600 reported pain in the shoulder at least once in the 12 months prior to the interview; 55 representing 18.3% (CI 14.3% to 23.1%) of the total, were men; and 89 representing 29.7% (CI 24.7% to 35%) of the total, were women.
It was possible to notice an association of pain in the shoulder with being female, older than 60 years, black and brown-skinned individuals, widowed or separated, and those who presented two or more reported morbidities (Table 1).
Concerning sedentary activities, there was no association between the studied variables and pain in the shoulder (Table 2).
There was no association of outcome with ergonomic variables ( Table 3).

DISCUSSION
In this study, the prevalence of shoulder pain was 24% lower than that found in Japan (30%) 1 , Netherlands (48%) 3 , and China (48.7%) 2 . A possible reason for these variations is that few studies have uniformly defined the clinical condition and anatomical area of the cervical and shoulder regions 1 .
Shoulder pain was associated with individuals 60 years or older, similar to the study in the Netherlands 3 , while in Japan it was associated with young adults 20 . The increased risk of shoulder pain in older adults is due to degenerative changes in muscles, tendons, ligaments and joints inherent in the natural process of senescence, chronic overload for the old aged worker and long-term exposure to occupational risk factors 3 .
Women were associated with the outcome, corroborating other studies 1,2,6 . These differences are due to greater exposure to static position, monotonous and repetitive tasks in daily life as well as domestic tasks, and to a lower pain threshold 1,2,6 . A Japanese study noted that women felt more stress and had more concerns, having negative effects on the muscles of the cervical and shoulder region. Although the sex distinction is still controversial, psychosocial and biological sex differences affect the occurrence of shoulder pain 1 .
The use of computer more than three times a week and work sitting have been associated with shoulder pain, according to some findings 2,6 . These factors associated with poor postural habits, ergonomically inadequate environments, and psychosocial factors lead to static contraction of shoulder muscles, contributing to the development of inflammatory processes in musculoskeletal structures 21 . However, a systematic review reported that sitting position is not a relevant risk factor for shoulder pain 22 .
Those works that require body tilting while standing remained associated with the outcome, similar to other studies 2,22 . Review studies have reported an association with occupational activities requiring upper limb strength, load lifting, repetitive shoulder and hand movements, using vibrating tools, hand raising above shoulder level, and the combination of these factors 23 .
The outcome was associated with two or more reported morbidities. The literature does not indicate the association of morbidities with shoulder pain. However, two studies on chronic lumbar musculoskeletal pain reported that pain is more commonly related in individuals with other associated chronic morbidities. The association of symptoms with diseases can be interpreted as cause or effect, due to the cross-sectional design of the studies.
This study had as limitations the non-collection of data regarding frequency, severity, intensity, and duration of pain, as well as psychological factors and how the shoulder discomfort affected and/or reduced occupational and daily life activities. Furthermore, the use of selfreport to measure the work conditions also limited the study outcomes. This study results are consistent with the literature, and it is consisted by a large number of interviews, representing its strong points.
This study will contribute as a reference for other epidemiological investigations and contribute to knowledge on the national prevalence estimative and its risk factors for systematic reviews and meta-analysis. Moreover, in Brazil the availability of national research data related to the result is scarce.
It was concluded that shoulder pain had a high prevalence and is associated with women, older aged, using the computer more than three times a week, performing occupational activities in the sitting position, incorrect sit-to-stand movement, and reporting two or more morbidities.