With regard to first-time TASDs, previous research has been limited to small cohort studies, case series and systematic reviews. The relevant background information supporting the need for research into the efficacy of management options for patients with a first-time TASD has been described. It highlighted that the use of traditional conservative management approaches after the initial reduction and joint immobilisation after a TASD seem to result in high rates of recurrent dislocation in some population groups.3,38–40 In younger patients, rates of recurrence as high as 92–96% have been reported.6 An incidence study of shoulder instability among athletes at a US military academy showed that 85% experienced a recurrent event within a 9-month period.7 A systematic review showed that there were some limited data to support primary surgery following a first-time TASD in young adults engaged in demanding physical activities (i.e. military personnel and athletes).5 A later systematic review also showed that in younger patients a significantly lower rate of recurrent instability was identified in the 2-year period following a first-time TASD for those having surgery than for those having no surgery (7% vs. 46%).8
Consequently, there appears to be some limited evidence for surgical intervention following a first-time TASD in younger and/or highly active patients. This research was, therefore, commissioned by the NIHR HTA programme as an ongoing research uncertainty, and probably with concern that this type of surgery is becoming more frequent after a first-time TASD, based on the positive suggestions of the lower-quality evidence described above.
Internal and external validation study
We conducted both an internal and an external validation assessment of the CPRD and found it an acceptable data set to identify and study shoulder dislocation patients. It provided a large, population-based, primary care cohort that is representative of the UK general population. We found the GP coding internally valid for shoulder dislocations and the incidence rates externally valid against data from Canada and the USA.
Our large UK population-based cohort of 16,763 patients aged 16–70 years during 1995–2015 allowed us to identify, for the first time (to our knowledge), the overall incidence rates in the UK. Most shoulder dislocations occurred in men (72%). The overall incidence rate in men was 40.4 per 100,000 person-years, and in women this was 15.5 per 100,000 person-years. The highest incidence was observed in 16- to 20-year-old men (80.5 per 100,000 person-years). Although this was similar to other world data (Canadian, US and Norwegian cohorts15,16,20), an unexpected finding was that the incidence in women aged > 50 years increased to 28.1 per 100,000 person-years among those aged 61–70 years.
Our new finding of the increasing incidence of shoulder dislocations among women aged > 50 years is of both interest and concern because the reasons for it are not known. Such injuries in more elderly people are usually associated with rotator cuff tears and fractures, with the subsequent loss of function as well as instability. However, this finding suggests that further work is now required to examine the reasons that may underpin this increased risk of shoulder dislocations in ageing women. Biological differences between ageing men and women in relation to joint proprioception, soft tissue tendon quality and protective muscle bulk and differences in the incidence of falls between men and women are all factors that could be examined. With an increasing ageing population, priority needs to be given to increasing the safety of the elderly to reduce falls, dislocations and fractures, as advocated by NICE;22 this is a new finding that warrants exploration.
Stage 2: propensity score analysis
One weakness identified during stage 1 of our study was that not all a priori risk factors were identifiable in the CPRD and the impact this would have on the stage 2 propensity score analysis was unclear because linkage to HES did not take place until stage 2 of this study. However, besides this weakness, once CPRD and HES linkage had taken place, a population-based cohort of 3759 patients diagnosed with a first-time TASD during 1 April 1997–26 April 2014 with 2 years of follow-up in England was produced.
Therefore, it was surprising that for a commissioned research question we could find only 156 patients in this large data set who had received surgical treatment within 6 months of a first-time TASD. On the one hand, this information is useful and informative and allows us to conclude that within the NHS early surgery after only one shoulder dislocation is uncommon. It also means that the study would be underpowered to demonstrate any real differences in re-dislocation rates between surgical and non-surgical treatments. Therefore, although the number of patients (n = 3759) included in the analysis was greater than the minimum number required for statistical power (n = 3065), the unexpected low number of NHS patients having surgery after one dislocation was a disappointing and surprising finding. It was also observed that a substantial number of patients in the young cohort of 16- to 35-year-olds had < 2 years of follow-up within the CPRD (n = 854), which may be related to them going away to university or finding jobs in different locations, resulting in the need to change their GP and this, in turn, resulted in a further loss of numbers. Overall, relatively few patients have surgery within 6 months of a first-time TASD in the NHS. This is probably a reflection of many GPs not referring patients with only one dislocation to secondary care and also due to NHS operative waiting times.
Combining this finding with the confounding risk factors not available in either the CPRD or HES had a large impact on the study. Although the overall finding from the propensity-score-matched analysis was that surgery within 6 months was slightly protective, it did not reach statistical significance (HR 0.88, 95% CI 0.58 to 1.35; p = 0.565) and the wide CI further indicates that this study was underpowered. With regard to the missing risk factors, the main disadvantage of using propensity-score-matching methods is that confounders for which no data are available result in a lack of adjustment. Other than age and sex, the risk factors recorded and available in the CPRD are considered less important risk factors for this particular condition. Other important factors identified in our expert survey, such as cause of shoulder dislocation, imaging findings of structural problems, anterior apprehension, occupation, sports played and level of sports, were not recorded in the observational data. Outcome data on ongoing instability symptoms without dislocation were also not available and, thus, only the hard outcome of re-dislocation could be used. This is another layer of potential confounding that could not be accounted for.
Sensitivity analysis
Based on the primary analysis findings, and in an attempt to maximise the use of this data set to further examine the commissioned question of surgery after a first-time TASD, a further sensitivity analysis was planned and approved by the HTA programme and ISAC. We looked at surgery within 12 months of a first-time TASD and increased the follow-up to up to 3 years. This produced a population-based cohort of 4613 patients diagnosed with a first-time TASD during 1 April 1997–31 March 2015 in England within CPRD-HES linked records. The overall finding from this propensity-score-matched analysis was that surgery within 12 months had a similar effect to non-surgical interventions (HR 1.17, 95% CI 0.88 to 1.55; p = 0.274) and did not seem to offer any additional benefit on whether or not a patient suffers a re-dislocation. However, the number of patients receiving surgery only increased to 342 (from 156 patients) and residual confounding was present.
There are two further observations to note from this analysis. First, re-dislocations in the surgical group seemed to occur later. This is unlikely to be due to any benefits of surgery wearing off but more likely to be related to the return to contact sports, which is a true test of stability. This is often delayed for > 6 months after surgery as part of the rehabilitation process. Second, the same process tends not to be in place for any non-operative patients; such patients may not even return to contact sports having decided to change their lifestyle instead of considering surgery. Recording such outcome metrics would be important for any future trials on the treatment of this condition. Although risk factors and residual confounding existed and it was not possible to reliably compare surgery with no surgery, it is still worth noting the 20% re-dislocation rate in the surgical cohort. The re-dislocation recurrence rate after surgery is probably higher than many surgeons and patients would expect and will help inform shared decision-making processes with patients.
Prediction models
Although some risk factor data were not available, it was still possible to construct prediction models, but these were limited. A non-surgical and a surgical model were developed and although the performances of the two models were markedly different, with a better performance in the surgery cohort, both models identified age and epilepsy as statistically significant predictors of re-dislocation. In the non-surgery cohort, sex (male) was also identified as a statistically significant predictor of re-dislocation. The modelling study has several strengths, as the models were developed using data routinely collected in electronic health-care records in primary care. Widely recommended statistical methodology was used to develop and evaluate the models, including the exploration of complex relationships in continuous measurements and bootstrapping techniques to internally validate them. The results indicate that using prediction models for this condition holds promise, but more risk factors are needed for more accurate prediction of outcome information for patients and surgeons.
Conclusions
- This study provides the first-time age- and sex-specific UK incidence rates for TASD, with most TASDs occurring in men, but with women aged > 50 years unexpectedly showing an increased incidence.
- Far fewer patients received surgery after a first-time TASD than expected, leading to an underpowered study.
- Surgery after a first-time TASD is not common in the NHS. Re-dislocation rates for surgical patients after a first-time TASD are higher than previously expected (at around 20%).
- A sensitivity analysis at 12 months suggests that there is little difference in re-dislocation rates between surgical patients and non-surgical patients, but important residual confounding risk factors were present and not recorded in NHS primary and secondary care databases.
- Missing risk factor data limited the value of the prediction modelling; however, age and epilepsy were identified as statistically significant predictors of re-dislocation.
A randomised controlled trial and/or a carefully constructed national shoulder dislocation registry documenting all appropriate risk factors and outcome metrics is needed to answer this commissioned research question reliably.