Author, date and country | Patient group | Study type (level of evidence) | Outcomes | Key results | Study Weaknesses |
---|---|---|---|---|---|
Beattie et al, 1986, UK | 111 consecutive patients (age range 16 -89) presenting to an Accident and Emergency Department with an anterior dislocation of the shoulder without a fractured neck of humerus nor neurovascular damage. | RCT with crossover design. Patients were randomised based on whether they presented on an odd or even date. Manipulation was performed by one of 4 investigators all of whom were "experienced in each technique". It is not stated whether analgesia +/- sedation had been given to the patients. On even dates a single attempt of Kocher's method was performed (N=55) and on odd dates a single attempt of Milch's method was performed (N=56). If the first of either technique was unsuccessful, a single second attempt was made using the alternative method. If the second attempt failed, then further manipulation was performed under GA. Reduction was confirmed both clinically and radiologically. | Successful reduction | Kocher's Technique slightly more successful than Milch's Technique (77% compared to 75% including 1st and 2nd attempts (i.e. combining pre and post cross over attempts)), however this failed to reach statistial significance. | 1. No sample size estimates were performed, perhaps leading to the results failing to reach statistical significance 2. There is no mention of ethical approval 3. The method of randomisation used may have introduced bias 4. The lack of information regarding analgesia and sedation makes it difficult to ascertain whether the results are valid and reliable 5. The trial was unblinded, but this seems to have been unavoidable 6. It is unclear whether the groups were comparable at baseline 7. The results are largely presented in the form of percentages which have been rounded to the nearest whole number. There is a discrepancy in the number of patients documented to have gone on to require GA (n = 11) and the number seeming to have had a failed reduction on the second attempt (n = 6.104) (based on calculations performed using the percentages given in the paper) 8. The statistical methods used have not been defined 9. The authors' conclusions are not based on statistically significant results |