Physical and psychosocial risk factors for musculoskeletal disorders in Brazilian and Italian nurses

As part of the international CUPID investigation, we compared physical and psychosocial risk factors for musculoskeletal disorders among nurses in Brazil and Italy. Using questionnaires, we collected information on musculoskeletal disorders and potential risk factors from 751 nurses employed in public hospitals. By fitting countryspecific multiple logistic regression models, we investigated the association of stressful physical activities and psychosocial characteristics with site-specific and multisite pain, and associated sickness absence. We found no clear relationship between low back pain and occupational lifting, but neck and shoulder pain were more common among nurses who reported prolonged work with the arms in an elevated position. After adjustment for potential confounding variables, pain in the low back, neck and shoulder, multisite pain, and sickness absence were all associated with somatizing tendency in both countries. Our findings support a role of somatizing tendency in predisposition to musculoskeletal disorders, acting as an important mediator of the individual response to triggering exposures, such as workload. Nursing Staff; Cross-cultural Comparison; Musculoskeletal Diseases; Absenteeism Introduction Musculoskeletal disorders are an important cause of morbidity in Western countries 1 where they have been a major focus for research. Moreover, in the last decade the body of evidence on musculoskeletal disorders has extended to also include epidemiological investigations conducted the so called “developing” countries 2,3,4. Many studies have highlighted the important role of stressful physical activities in the generation and progression of musculoskeletal disorders. At the same time, psychosocial risk factors, such as job satisfaction, somatizing tendency and mood, are receiving growing attention as determinants of musculoskeletal disorders, and appear at least equally important 5,6,7,8. In addition, some studies have started to explore the role of culturally determined health beliefs in generating and maintaining musculoskeletal symptoms. These studies have suggested marked variations in the prevalence of common musculoskeletal complaints and associated disability among workers carrying out similar jobs but in dissimilar settings (different countries) 9,10. It is hypothesized that in many cases, chronic musculoskeletal symptoms and disability may be “psychologically mediated responses to triggering exposures” 11 (p. 281) conditioned by individual characteristics and cultural circumstances. Nurses employed in hospitals are particularly liable to work-related musculoskeletal disorders: 1632 ARTIGO ARTICLE RISK FACTORS FOR MUSCULOSKELETAL DISORDERS IN NURSES 1633 Cad. Saúde Pública, Rio de Janeiro, 28(9):1632-1642, set, 2012 their work frequently involves heavy lifting, often in awkward postures, and sometimes entails forceful movements of the upper limbs. Low back, neck and shoulder pain have been shown to be highly prevalent among nurses 12. In order to study the psychosocial and cultural influences on musculoskeletal disorders, we recruited two populations of Brazilian and Italian nurses and nursing technicians (called nurses as a whole), as part of the international CUPID (Cultural and Psychosocial Influences on Disability) study. Our aim was to compare the prevalence of musculoskeletal disorders among nurses in the two locations, and their association with physical and psychosocial risk factors.


Introduction
Musculoskeletal disorders are an important cause of morbidity in Western countries 1 where they have been a major focus for research.Moreover, in the last decade the body of evidence on musculoskeletal disorders has extended to also include epidemiological investigations conducted the so called "developing" countries 2,3,4 .
Many studies have highlighted the important role of stressful physical activities in the generation and progression of musculoskeletal disorders.At the same time, psychosocial risk factors, such as job satisfaction, somatizing tendency and mood, are receiving growing attention as determinants of musculoskeletal disorders, and appear at least equally important 5,6,7,8 .In addition, some studies have started to explore the role of culturally determined health beliefs in generating and maintaining musculoskeletal symptoms.These studies have suggested marked variations in the prevalence of common musculoskeletal complaints and associated disability among workers carrying out similar jobs but in dissimilar settings (different countries) 9,10 .It is hypothesized that in many cases, chronic musculoskeletal symptoms and disability may be "psychologically mediated responses to triggering exposures" 11 (p.281) conditioned by individual characteristics and cultural circumstances.
Nurses employed in hospitals are particularly liable to work-related musculoskeletal disorders: their work frequently involves heavy lifting, often in awkward postures, and sometimes entails forceful movements of the upper limbs.Low back, neck and shoulder pain have been shown to be highly prevalent among nurses 12 .
In order to study the psychosocial and cultural influences on musculoskeletal disorders, we recruited two populations of Brazilian and Italian nurses and nursing technicians (called nurses as a whole), as part of the international CUPID (Cultural and Psychosocial Influences on Disability) study.
Our aim was to compare the prevalence of musculoskeletal disorders among nurses in the two locations, and their association with physical and psychosocial risk factors.

Methods
We conducted a cross-sectional survey focusing on nurses from large public hospitals in Brazil and Italy.Between May 2008 and March 2010, we recruited nurses who were employed in medical wards at São Paulo University Hospital (Brazil), and Milan and Varese University Hospitals (Italy), here on called Brazilian and Italian nurses, respectively, for the sake of brevity.To be included in the study, subjects had to have been employed in their current job for at least 1 year.
The study protocol was approved by each Institutional Review Board, and written informed consent was obtained from all participants.Each subject completed a self-administered questionnaire in their native language.The questionnaire for the CUPID study was originally compiled in English, and translated into Portuguese and Italian.To check the accuracy of the translations, each questionnaire was independently back-translated to English.The back-translated versions were then compared to the original to identify any inconsistencies, and re-submitted to the central coordinator of the CUPID study (D.C.) who suggested adjustments.In addition, as described by a previous study 3 , the Portuguese version of the questionnaire was pre-tested on a sample of nurses before applying it to the main study.
Questions covered aspects such as demographic characteristics, hours worked per week, duration of employment, whether the nurse's job involved specific physical activities in an average working day, job satisfaction, tendency to somatize, mental health (mood), pain at specific anatomical sites, and related disability and sickness absence.

Exposure assessment
We categorized hours worked per week using a cut-off of 38 hours (a full-time working week).We assessed job satisfaction by asking participants directly how satisfied they were with their job as a whole and grouping answers to form a dichotomous variable (satisfied vs. dissatisfied).To measure somatizing tendency, we used elements of the somatic subscale of the Brief Symptom Inventory 13 , asking about distress caused by nausea, faintness, dizziness, weakness, numbness, chest pain and breathing difficulties in the previous week.Participants were then classified according to the number of these symptoms (zero, one, two or more) causing at least moderate concern.We assessed mental health using questions from the relevant section of the Short Form-36 questionnaire 14 .The sum of scores resulting from individual questions was grouped into approximate tertiles of their distribution in the whole study sample (poor, intermediate or good mental health).
The CUPID questionnaire focused on six anatomical sites: low back, neck, shoulder, elbow, wrist/hand and knee.For each site, we identified one or more stressful physical activities in an average working day including: lifting a weight of ≥ 25kg by hand (low back) and work with hands above shoulder height for ≥ 1 hour (neck and shoulders).

Outcome definition
Pain at different anatomical sites was assessed by means of specific questions regarding location and duration of pain, whether the pain had made one or more everyday activity (such as getting dressed, doing normal household jobs, etc.) difficult or impossible, and whether it had led to absence from work; to avoid misunderstanding regarding pain location, each anatomical site was depicted by an image.The simplest outcome measures were the presence or absence of pain in the past month at each of the six anatomical sites of interest.Pain was then defined as "disabling" if it had made at least one of the specified everyday activities difficult or impossible in the past month.We also investigated whether pain had occurred in three or more sites in the past month (multisite pain), and whether sickness absence had occurred in the past year because of musculoskeletal pain.When considering regional pain, we focused only on those anatomical sites which are most likely to be stressed by the typical activities of nurses employed in hospital wards: low back, neck, and shoulder.However, when investigating multisite pain and sickness absence, all anatomical sites were taken into account.

Statistical analysis
We first compared the occupational and psychosocial characteristics of nurses in the two locations, by means of descriptive statistics, Pearson's chi-squared test, and Mann-Whitney U test for ordinal variables.We then fitted location-specific multiple logistic regression models to assess the association of risk factors with pain, disabling pain, multisite pain and sickness absence; for categorical variables with more than two levels, a test for trend was performed.All statistical tests were two-sided where; a p-value < 0.05 was considered statistically significant.The statistical analysis was performed using Stata/MP 11.1 (Stata Corp., College Station, USA).

Results
A total of 969 nurses were invited to take part in the study.Questionnaires were returned by 195 Brazilian nurses (participation rate: 96%) and 585 Italian nurses (participation rate: 76%).However, we excluded two Brazilian and 27 Italian nurses because they had been employed in their current job for less than one year.Thus, our analysis was based on 751 subjects.
Table 1 summarizes the main characteristics of the study sample by location.Gender distribution was similar in the two locations (84% females) but Italian nurses were younger (50% aged < 40 vs. 39%) with a shorter duration of employment.Approximately 72% of participants reported an average working week of 38 hours or less, with a higher proportion of Italian nurses (26% vs. 19%) working more than 38 hours/week.Both lifting a weight of 25kg or more and working with the hands above shoulder height for an hour or longer were more frequent among Italian participants.Approximately 92% of Brazilian nurses declared they were satisfied with their current job, as compared with 83% of Italian nurses.In both locations, more than 30% of the study participants reported two or more somatic symptoms causing at least moderate concern (somatizing tendency).Mood was poorer among Brazilian nurses, with more than 65% of the participants having intermediate or poor mental health.The corresponding proportion among Italian nurses was 55%.
Table 2 shows the prevalence of regional pain among participating nurses by location.Rates of low back pain in the past month were similar (45% in São Paulo vs. 49% in Milan/Varese), but there was a higher prevalence of related absence in the past year in the Italian cities. Italian nurses reported more neck pain, although the difference was not statistically significant.Brazilian nurses reported more shoulder problems, with significant differences for shoulder pain in the past month (42% in São Paulo vs. 33% in Milan/ Varese, p = 0.04) and shoulder pain causing absence in the past year (12% in Brazil vs. 7% in Italy, p = 0.04).The frequency of multisite pain was different in the two locations, with a higher proportion of Brazilian nurses reporting ≥ 3 painful anatomical sites in the past month (42% vs. 30%); however, the percentages of participants reporting no painful sites at all were very similar (23% in São Paulo vs. 25% in Milan/Varese).
Table 3 shows the results of the multiple logistic regression analyses, exploring associations between the main investigated risk factors and pain at different anatomical sites in the past month.The reference category for each of the risk estimates in this table was no pain at the investigated site during the past month.All regression models included sex, age, hours worked per week, site-specific stressful physical activity, job satisfaction, somatizing tendency and mood; associations are summarized by odds ratios (OR) and corresponding 95% confidence intervals (95%CI).With regard to low back pain, our analysis showed a lower risk in men than in women, which reached statistical significance among Italian nurses (OR = 0.51, p = 0.02).Increasing age was clearly associated with higher risk of low back pain in the Italian nurses and a similar pattern was apparent in the Brazilian nurses, although none of the OR was statistically significant.Working for more than 38 hours per week carried an increased risk only among Italian nurses (OR = 1.83, p = 0.01).No significant association was found between low back painand lifting a weight of 25kg or more in an average working day, although OR in both locations were greater than one.Job dissatisfaction also carried an increased risk (OR = 1.52 in São Paulo and OR = 2.27 in Milan/Varese), but this reached statistical significance only in Italy (p = 0.002).We found a strong relationship between somatizing tendency and risk of low back pain, with OR ranging from 1.49 among Brazilian nurses reporting one symptom causing at least moderate concern in the past week to 2.99 among Italian nurses reporting two or more distressing symptoms: in both locations associations were statistically significant (p-trend = 0.02, and p-trend < 0.001 in São Paulo and Milan/Varese respectively).No association was found with mood.
As regards neck pain in the past month, associations with gender and age were similar to  those for low back pain.No association was observed with hours worked per week, but stressful physical activity (working with hands above shoulder height for one hour or longer) was associated with a significantly elevated risk in both Brazilian and Italian nurses (OR = 5.45, p = 0.01 and OR = 2.11, p = 0.001, respectively).No significant association was observed with job satisfaction, although in both locations OR exceeded one.A positive association with somatizing tendency was also observed for neck pain, with a clear trend in risk among the Italian nurses (p < 0.001).As for low back pain, no association was observed with poor mental health.Shoulder pain was also more common at older ages.In addition, there was a positive association with stressful physical activity (work with hands above shoulder height for ≥ 1 hour), with similar OR in the two locations (1.72, p = 0.4 in São Paulo and 1.89, p = 0.01 in Milan/Varese).Job dissatisfaction was strongly associated with shoulder pain in São Paulo (OR = 8.06, p = 0.01).A positive association was also shown in Milan/Varese, but it was not statistically significant (OR = 1.53, p = 0.1).Somatizing tendency was a risk factor for shoulder pain and for pain at other sites, with OR ranging from 1.21 (Italian nurses reporting one symptom causing at least moderate concern in the past week) to 4.78 (Brazilian nurses reporting two or more distressing symptoms).
Associations with disabling pain in the past month were generally similar to those for any pain at the same anatomical site (Table 4).Thus, age, hours worked per week, job dissatisfaction and somatizing tendency were associated with disabling low back pain.Working with hands above shoulder height for an hour or longer in an average working day was significantly associated with disabling neck pain in both São Paulo (OR = 5.44, p = 0.01) and Milan/Varese (OR = 2.17, p = 0.001).In both locations, somatizing tendency was positively associated with disabling neck pain (p = 0.02 and p < 0.001 for São Paulo and Milan/ Varese, respectively) and disabling shoulder pain (p < 0.001 for both locations).A 64% increased risk of disabling shoulder pain was observed among Italian nurses working more than 38 hours per week.It was also shown that job dissatisfaction carried an OR of 5.59 (p = 0.05) for disabling shoulder pain among Brazilian nurses.
Further analysis (Table 5) confirmed somatizing tendency to be a strong risk factor for reporting pain at three or more sites in the past month (reference category: pain at two or fewer sites) in both Brazilian (OR = 2.37 for one symptom causing at least moderate concern, OR = 3.15 for two or more symptoms, p-trend = 0.004) and Italian (OR = 1.89 for one symptom, OR = 3.51 for two or more symptoms, p-trend < 0.001) nurses.Positive associations were also found between somatizing tendency and sickness absence in the past year, both in São Paulo (OR = 1.91 for one symptom, OR = 3.14 for two or more symptoms, p-trend = 0.01) and in Milan/Varese (OR = 1.55 for one symptom, OR = 2.32 for two or more symptoms, p-trend = 0.001).

Discussion
Our study evaluated musculoskeletal disorder prevalence in two occupational groups from different socio-cultural backgrounds, both performing very similar job tasks (nursing staff from medical wards in large public hospitals in Brazil and Italy) and compared, across different loca- tions, the relation of risk factors to pain, disability and sickness absence.
In our population, musculoskeletal disorder prevalence tended to be higher than that previously reported from similar occupational settings in other countries 15 .
Despite possible differences in working conditions, systems of remuneration and worker beliefs, no major inconsistencies were observed between the Italian and Brazilian nurses studied with respect to low back pain prevalence or related disability and sickness absence.
After adjustment for the effect of other individual and psychosocial risk factors, physical activity (lifting a weight of ≥ 25kg by hand) was not significantly associated with low back pain or disabling low back pain in either Brazilian or Italian nurses.
The prevalence of neck pain and shoulder pain was similar to the observed prevalence of low back pain in both locations.An increased risk of neck and shoulder pain was observed for subjects working with the hands above shoulder height for at least one hour/day.
Somatizing tendency appeared to be a relevant risk factor for all the investigated outcomes, particularly for disabling pain, multisite pain and sickness absence.Despite possible differences in cultural and social backgrounds, these associations were seen in both locations and are consistent with findings in other studies 15,16 .
Major strengths of our study are: a relatively high participation rate, comparison of the same occupational group in two locations characterized by different social and cultural backgrounds, and the ability to evaluate not only pain prevalence but also -as suggested by recent investigations 17,18 -its consequences (disability and sickness absence).
The participation rate was 96% among Brazilian and 76% among Italian nurses, being higher or just slightly lower than response rates obtained elsewhere in the CUPID study 12,15 .Important response bias is therefore unlikely.While non-responders may have differed somewhat from responders in their prevalence of musculoskeletal symptoms, it seems unlikely that the differences were so great as to seriously bias prevalence estimates.Moreover, when considering associations between musculoskeletal symptoms and risk factors, important bias is even more unlikely.These response rates were achieved while ensuring the anonymity of participants and in strict collaboration with the occupational physicians at the investigated hospitals.
Applying the same study protocol and standardized measurements of exposures and outcomes, we were able to investigate and compare the associations of prevalent musculoskeletal disorders with potential risk factors in São Paulo (located in a rapidly developing country in which musculoskeletal disorders in occupational settings were extensively investigated only recently 19,20 , usually without taking into account psychosocial risk factors) and in Milan and Varese (both located in a country characterized by a higher median income and a more stable economic environment).Moreover, the relation of somatizing tendency to musculoskeletal disorders had not previously been investigated in Brazil.
The main limitations of our study arise from its cross-sectional design: as is well known, the contemporary collection of data on both risk factors and health limits conclusions drawn about causal relationships.We have therefore presented our findings with caution, without interpreting them as causal relationships, and referring to "related" or "'associated"' factors 21 .The direction of cause and effect in cross-sectional associations with musculoskeletal disorders is uncertain 22 .In our study, the lack of significant association between physical activity and low back pain may be a consequence of healthy worker selection, arising because nurses with musculoskeletal disorders tend to move (or be moved) to other hospital units where there is less frequent lifting and moving of patients.In addition, the observed association between job satisfaction and pain may be influenced by a tendency for nurses to perceive a higher workload and feel less satisfied if they are experiencing frequent musculoskeletal pain (reverse causation).
Reverse causation may also have played a role in the observed association between somatization and pain prevalence.Workers with musculoskeletal disorders might be prone to describe their "general health" more negatively.The items involved in the somatizing tendency measurement included both nonspecific symptoms (such as dizziness, chest pain, nausea and breathing difficulties), and two neurological symptoms that could in some circumstances arise from musculoskeletal disease ("feeling weak in parts of your body" and "numbness or tingling in parts of your body").When we repeated analyses excluding these two questions, the association of somatizing tendency with low back pain (São Paulo), shoulder pain (both locations), multisite pain (São Paulo) and sickness absence (Milan/Varese) lost their statistical significance, but odds ratios were generally only slightly reduced and in other cases remained statistically significant (data not shown).
Our study relied on self-reported information, which may have led to misclassification of some exposures.For example, a worker who was currently suffering from musculoskeletal pain might be more aware of stressful physical activities and report them more frequently.The effect of any such misclassification would be to inflate OR, and it would not explain the absence of significant associations between low back pain and lifting.
To address some of the above mentioned limitations we are planning to follow up workers for at least one year and to re-measure musculoskeletal disorders and other outcomes at the end of this period.This will allow us to prospectively evaluate the effect of the evaluated risk factors on the risk of developing new musculoskeletal disorders and the persistence of musculoskeletal disorders already present at baseline.
In conclusion, our findings support a possible role of somatizing tendency in predisposing to musculoskeletal disorders.The influence of psychosocial and cultural characteristics on musculoskeletal disorders prevalence is well described in the earlier literature 5,23 , suggesting that they are important mediators of the individual response to triggering exposures (such as workload) 11 .Among such characteristics, somatizing tendency is likely to play an important role, with effects across different cultural environments.

Recursos Humanos em Enfermagem; Comparação Transcultural; Doenças Musculoesqueléticas; Absenteísmo
Cad. Saúde Pública, Rio de Janeiro, 28(9):1632-1642, set, 2012 Contributors M. Carugno and M. Bonzini contributed with the conception and design, acquisition analysis and interpretation of data, drafting the article and final approval of the version to be published.A. C. Pesatori and D. Coggon contributed with the conception and design, interpretation of data, critical revision of the article and final approval of the version to be published.M. M. Ferrario contributed with the conception and design, acquisition and interpretation of data, critical revision of the article and final approval of the version to be published.A. L. Ferrari, F. J. Silva, and A. C. Martins contributed with the interpretation of data, critical revision of the article and final approval of the version to be published.V. E. A. Felli contributed with the conception and design, acquisition and interpretation of data, drafting and critical revision of the article and final approval of the version to be published.
** Percentages may not add up to 100.0 because of rounding; *** Number of symptoms in past week causing at least moderate concern.

Table 2
Prevalence of regional pain among participating nurses by location * , **.
* p = chi-squared test; for "Number of painful sites in past month", Mann-Whitney U test was performed; ** Percentages may not add up to 100.0 because of rounding.

Table 3
Associations between risk factors and pain in the past month by location *.
* Multivariable logistic regression models adjusted for all the variables presented in the table; for categorical variables with more than two levels, a test for trend was performed; ** Stressful occupational activity in an average working day defi ned as lifting a weight of ≥ 25kg by hand (low back) or work with the hands above shoulder height for ≥ 1 hour (neck and shoulder); *** Number of symptoms in past week causing at least moderate concern.

Table 4
Associations between risk factors and disabling pain in the past month by location *.Multivariable logistic regression models adjusted for all variables presented in the table; for categorical variables with more than two levels, a test for trend was performed; ** Stressful occupational activity in an average working day defi ned as lifting a weight of ≥ 25kg by hand (low back) or work with the hands above shoulder height for ≥ 1 hour (neck and shoulder); * *** Number of symptoms in past week causing at least moderate concern.

Table 5
Associations of selected risk factors with multisite pain and sickness absence by location.
# Number of symptoms in past week causing at least moderate concern.