Population-based consultation patterns in patients with shoulder pain diagnoses

Musculoskeletal disorders are common in industrial countries [1, 2] and bring enormous costs to the society. In the European Union the cost of treatment and lost productivity is estimated to 0.5-2% of the gross domestic product [3, 4] and in 2000 in the United States, the direct cost for treatment of shoulder dysfunction was estimated to $7 billion [5]. There was approximately 3.75 million working days per year lost in the United Kingdom (2008–2009) due to musculoskeletal problems [6]. In Sweden musculoskeletal problems represent about one third of all sick-leave and make up for the majority of all long-term sick leave [7, 8]. Besides the economical burden, shoulder pain causes great individual harm and influences both work and private life.

The shoulder has been reported as the third most common site of musculoskeletal pain after low back pain [9] and knee pain [10]. In the general population a point prevalence of 7-26%, and one-year prevalence of 7-47% of self-reported shoulder pain are reported [3, 11‐14] while the annual incidence is estimated to be around 1-2% [15]. Among patients in primary care the annual consultation prevalence has been estimated to range from 2-10% of the population [15‐17] and incidence from 11-30/1000 person-years [14‐16, 18]. Besides individual risk factors, such as age, arthritis, obesity, diabetes and thyroid disease [3, 19‐21], a strong relationship between working conditions and shoulder disorders has been reported in several studies [11, 22, 23]. Also, regional pain in the shoulder can evolve into more generalized pain syndromes [24]. The gender impact on shoulder disorders is not clear and is a topic of discussion [25, 26].

Considering the tremendous burden of shoulder conditions on society, an up-to-date estimate of the burden on the health care system is important. Further, the prognosis, in terms of repeated consultations, is also of interest. Many previous studies are based on questionnaires often including mild and transient self-reported symptoms [15]. However, only about 20-50% of subjects with shoulder symptoms are likely to ever consult a doctor for their problem [15, 16]. To assess such an estimate, registers of health care may be used. Since shoulder complaints are typically handled in primary care, information on both in-patient and out-patient data is required. The Skåne Health Care Register (SHCR), which covers the population in the southernmost part of Sweden, encompasses such data with diagnostic codes by medical doctors. Enabled by SHCR, the aim of this study was to gain new insights of the consultation patterns of doctor-diagnosed shoulder disorders by age and gender.

We used population-based health care register data from Skåne County, the most southern county of Sweden (population Dec 31, 2005: women 593,569; men 575,895; total 1,169,464). The study was approved by Lund University institutional review board.

The annual consultation prevalence for shoulder pain diagnoses was about 1% of the adult population and similar in both women and men. The new onset consultation rate was 80/10 000 per year for women and 74/10 000 per year for men. The consultations for shoulder pain diagnoses increased by age and peaked for women at age 50–59 and for men at age 60–69, where the incidence was much higher for both genders, compared to the whole working age group. Our study shows a somewhat lower annual consultation prevalence and new onset consultation rate leading to consultation in general practice than previous studies from UK and the Netherlands that suggest a prevalence and incidence of 236-480/10 000 and 112-300/10 000 [14‐18, 30] respectively. Possible explanations could include, but not be limited to, differences in study design (e.g. case criteria), health care systems and/or the more frequent use of symptomatic codes. Linsell et al. for example, also included acute trauma diagnoses and, Bot et al. included a general diagnostic code for “shoulder complaints”. Dorrestijn et al. report that in only 14% a specific diagnosis was recoded [30]. When taking this into account, the occurrence of shoulder pain conditions in Sweden seems to be about similar as in the United Kingdom and Holland.

Many shoulder pain diagnoses are associated with increasing age and a general degeneration of the body. In accordance with previous studies, new onset consultation rate increased by age, showing a peak in age 50–60 [14‐16]. In fact the shape of the curve for new onset consultation rate had a similar shape in the report from Bot et al. compared to our Figure 2, though we present results for women and men separately. There was an increase in the new onset consultation rate after entering working life and a leveling off after retirement age (in Sweden at age 65.

In concordance with previous studies, tendinitis, bursitis and impingement syndrome, were the most frequent group of diagnoses for both women and men [5, 18, 22] Van der Wind et al. report that “subacromial syndrome” was the most frequently diagnosed disorder , in particular rotator cuff tendonitis (29%) [18].

Most patients with a new onset shoulder pain diagnosis consulted a doctor only once, which is in accordance with previous studies where 50% of new consulters only visited their GP once [15, 16] and only few continued consulting beyond two years, in particularly the elderly [15]. We chose to exclude subjects with a consultation free period of more than one year although some of these may still have ongoing pain, as well as being at higher risk for relapses. Not consulting does naturally not equal being free of symptoms. A history of persistent shoulder complaints has shown to be a good predictor for slow recovery over time or recurrence of complaints [31, 32] and about half of patients in primary care still reported complaints after 12 months [33] or experienced recurrent episodes during an 18-month follow up [34].

Some previous studies have shown that women have a higher risk of upper extremity disorders, compared to men (age and occupation adjusted) [25, 35, 36]. A systematic review of several databases found that women more often than men had neck-shoulder complaints with job tasks involving the same arm postures as men, while for job performances with hand-arm vibration men dominated [26]. Some studies indicate that gender differences in response to physical work exposure may reflect gender segregation in work and differences in pinch and lifting capacity [37, 38]. Our study did not show any substantial differences between new onset consultation rate of shoulder pain diagnoses among women and men at large, which is in accordance with the result from Linsell et al[15]. Other studies did find a higher incidence among women [14, 17, 18]. Our results showed, however, that women reached their peak of annual consultation prevalence as well as new onset consultation rate earlier in their lifespan than did men. Notably, within the group of shoulder pain diagnoses, men consulted more frequently due to impingement syndrome compared to women, while adhesive capsulitis, which is not typically associated with occupational conditions [39], was more often diagnosed among women. Possibly the patients sex could influence the choice of diagnosis, so that it is more likely that the doctor will give a diagnosis of frozen shoulder to a woman than a man, since it is known that this condition is more common among women [40].

Prevalence and new onset consultation rate were about equal in women and men. The rates increased from entering working life and decreased from retirement age. Shoulder pain diagnoses with a possible relationship to adverse ergonomics were the most frequent.

The Swedish Research Council. Faculty of Medicine, Lund University, The Swedish Social Insurance Agency, Region Skåne and the County Councils of Southern Sweden.

Among patients in primary care the annual consultation prevalence has been reported to range from 2-10% of the population and incidence from 11-30/1000 person-years. However detailed information on for example age and sex characteristics and the number of repeat consultations is scarce.

An up-to-date Swedish estimate show an annual consultation prevalence of 1%, in men and women, and a new onset consultation rate of 80/10 000 women and 74/10 000 men.

Both measures increased by age up to retirement age, and then decreased.