Compression Fractures Welcome UNC 椎体压缩性骨折欢迎北课件.ppt

Complications… ?adjacent vertebral fractures (probable) Most studies show increased risk Problem: patients with compression fractures have high probability of future fractures (25%/year) Confounding: Those with worse disease more likely to have VP/KP and more likely to have future fractures Concerns about the Validity of most recently reported studies… Outpatients (inpatients may have more severe pain) Patients received 4 weeks of medical treatment – patients on average had 9-16 weeks of symptoms in the 2 recent VP studies (compared to 6 weeks for the Lancet KP study) Counter: no difference in subgroup analysis between patients with less than or more than 6 weeks of symptoms Jewitt Brace The RCT… 1930: Sollman suggests approach to problem of investigator bias: use of blinded observer and a placebo control 1932-1937: Harry Gold at Cornell refines the double blind method and use of placebo 1935: Ronald Fisher’s “The Design of Experiments” argues for use of strictly randomized allocation The RCT Randomization made test groups more comparable and “ethical” 1947: limited supply of streptomycin for British patients, Bradford Hill in the BMJ pushed for studies with a randomized design: “precluded the biases introduced by our personal idiosyncracies, consciously or unconsciusly applied, or lack of judgment” RCT… 1960s: increase value on statistical evidence in interpreting evidence 1990s: Evidence Based Medicine… Won’t get fooled again… Hip protectors and decreased hip fractures… Estrogen use in postmenopausal women decreases the risk of CAD (women on estrogens live 1.5 years longer than those not…) Early coronary intervention must be good for patients with diabetes and evidence for significant but asymptomatic coronary disease on angiography Maybe trephination…. Problems with prior studies looking at VP and KP Not blinded Bias on part of investigators (evidence that it“works”) Bias of participants (advertised “evidence” that this works) Underestimated placebo