Quasi-experimental study designs series-paper 7: assessing the assumptions.

Quasi-experimental designs are gaining popularity in epidemiology and health systems research-in particular for the evaluation of health care practice, programs, and policy-because they allow strong causal inferences without randomized controlled experiments. We describe the concepts underlying five important quasi-experimental designs: Instrumental Variables, Regression Discontinuity, Interrupted Time Series, Fixed Effects, and Difference-in-Differences designs. We illustrate each of the designs with an example from health research. We then describe the assumptions required for each of the designs to ensure valid causal inference and discuss the tests available to examine the assumptions.

Publication selection in health policy research: the winner's curse hypothesis.

There is a widely discussed problem of publication bias in medical and health services research. Where quantitative effects form the basis of a publication a 'winner's curse' curse may apply. This phenomenon may occur as prospective authors of research papers compete by reporting 'more extreme and spectacular results' in order to increase the chances of their paper being accepted for publication. This paper examines this phenomenon using quantitative findings on income and price elasticities as reported in health economics research. We find robust statistical evidence that higher-impact journals preferentially report larger empirical estimates of these elasticities. That is, we find robust evidence of a winner's curse hypothesis contributing to the existence of publication bias found in both the income and the price elasticities of health care and drugs, as well as value of life research.

Repeat use of human recombinant bone morphogenetic protein-2 for second level lumbar arthrodesis.

STUDY DESIGN: Prospective randomized controlled animal model. OBJECTIVE: The purpose of this study is to determine whether the readministration of human recombinant bone morphogenetic protein-2 (rhBMP-2) induces an immune response and inhibits successful fusion in repeat posterolateral spinal surgery. SUMMARY OF BACKGROUND DATA: Little research has been performed on the effectiveness or immunoreactivity of rhBMP-2 (Infuse, Medtronic, Memphis, TN) in the context of its reuse in posterolateral fusion spinal surgery at adjacent levels. METHODS: A total of 34 New Zealand White rabbits underwent posterior intertransverse process fusion with the use of rhBMP-2 delivered on an absorbable collagen sponge (rhBMP-2/ACS). Two rabbits were killed early leaving 32 total rabbits. Serologic studies (Type I bovine collagen and rhBMP-2 antibodies) were obtained at 2-week intervals throughout the experiment. At 10 weeks, posteroanterior radiographs confirmed solid fusion masses in all rabbits. The 32 rabbits were randomly separated into 2 groups of 16, and each group underwent an adjacent level, bilateral intertransverse process fusion with either rhBMP-2/ACS or iliac crest. RESULTS: There was no statistical difference in fusion rates with repeat use of rhBMP-2 (n = 15/16, 94%) or iliac crest (n = 11/16, 69%) (P = 0.17) at the adjacent level. Four rabbits (n = 4/32, 13%) developed rhBMP-2 antibodies. Of these 4 rabbits, 1 developed anti-rhBMP antibodies after the first exposure and 3 developed antibodies after the second surgery. Eight rabbits (n = 8/32, 25%) developed collagen antibodies with 7 rabbits developing antibodies after the first exposure and 1 rabbit developing antibodies after the second exposure. The development of antibodies did not effect fusion rates. No rabbit demonstrated evidence of a systemic or anaphylactic reaction to repeat exposure to rhBMP-2. CONCLUSION: rhBMP-2 appears to be successful in promoting intertransverse fusions when used in both primary and repeat fusion environments. The infrequent development of antibodies to rhBMP-2 after re-exposure occurs without a predictable time course suggesting that host immunologic variation may play a role. This animal study would tend to support early clinical data emerging on the reuse of BMP-2 for lumbar fusion, suggesting an acceptable fusion rate without a high incidence of complications.

Percutaneous treatment of vertebral compression fractures: a meta-analysis of complications.

STUDY DESIGN AND OBJECTIVE: This study performs a meta-analysis to compare complication rates from vertebroplasty (VP) and kyphoplasty (KP). SUMMARY OF BACKGROUND DATA: Recently, the development of VP and balloon KP has been shown to provide symptomatic relief and restoration of sagittal alignment of vertebral compression fractures refractory to medical therapy. Complications in treatment of vertebral compression fractures are rare, however can be potentially devastating. Fortunately, clinical sequelae are rare; however, severe clinical complications from cement extravasation have been reported. METHODS: Using PubMed and Ovid, we performed a literature search for "kyphoplasty," "vertebroplasty," and "vertebral augmentation." This search was performed in December 2006. Case reports and reports not available in English were excluded. We categorized complications in 3 categories: (1) procedure-related complications, (2) medical complications, and (3) new vertebral fracture. Cement leakage, asymptomatic and symptomatic, and its locations were recorded. We performed a meta-analysis of complications of all studies. We then repeated the meta-analysis examining only prospective studies. We then used proportion analysis to determine statistical significance. We defined statistical significance as a P value less than 0.05. RESULTS: We identified 121 reports of KP and/or VP that specifically addressed complications. Of these studies, 33 addressed KP and 82 addressed VP (6 reports addressed complications of both). There were 29 reports in which the data appeared to be collected prospectively. Of these, 9 addressed KP and 21 addressed VP.VP was found to have a significantly increased rate of procedure-related complications than KP in the analysis of all studies and only prospective studies. VP also appears to have a significantly higher rate of symptomatic and asymptomatic cement leakage than KP (P < 0.05). The incidence of medical complications was significantly higher in the KP procedure; however, this difference was not observed when analyzing only prospective studies. The incidence of new fracture was significantly higher in the VP procedure; however, this was not observed when analyzing only prospective studies. CONCLUSION: VP and KP are 2 minimally invasive procedures that have been shown to be effective in the treatment of symptomatic vertebral compression fractures. Although the incidence of adverse events for both VP and KP are low, it appears that VP is associated with a statistically significant increased rate of procedure-related complications and cement extravasation (symptomatic and asymptomatic). Future prospective studies with large patient enrollment will be needed to further validate the finding of this meta-analysis.

Assessment of regression-based methods to adjust for publication bias through a comprehensive simulation study.

BACKGROUND: In meta-analysis, the presence of funnel plot asymmetry is attributed to publication or other small-study effects, which causes larger effects to be observed in the smaller studies. This issue potentially mean inappropriate conclusions are drawn from a meta-analysis. If meta-analysis is to be used to inform decision-making, a reliable way to adjust pooled estimates for potential funnel plot asymmetry is required. METHODS: A comprehensive simulation study is presented to assess the performance of different adjustment methods including the novel application of several regression-based methods (which are commonly applied to detect publication bias rather than adjust for it) and the popular Trim & Fill algorithm. Meta-analyses with binary outcomes, analysed on the log odds ratio scale, were simulated by considering scenarios with and without i) publication bias and; ii) heterogeneity. Publication bias was induced through two underlying mechanisms assuming the probability of publication depends on i) the study effect size; or ii) the p-value. RESULTS: The performance of all methods tended to worsen as unexplained heterogeneity increased and the number of studies in the meta-analysis decreased. Applying the methods conditional on an initial test for the presence of funnel plot asymmetry generally provided poorer performance than the unconditional use of the adjustment method. Several of the regression based methods consistently outperformed the Trim & Fill estimators. CONCLUSION: Regression-based adjustments for publication bias and other small study effects are easy to conduct and outperformed more established methods over a wide range of simulation scenarios.

The effect of stem design on end-of-stem pain in revision total knee arthroplasty.

A series of 143 revision total knee arthroplasties was evaluated using a Knee Society clinical score, radiographic assessment, and satisfaction questionnaire. Patients completed a pain drawing showing the location and severity of pain they experienced at 2 years follow-up. Two stem designs were used: a solid, fluted cobalt chrome stem (group I) and a slotted titanium stem (group II). End-of-stem pain was associated with solid cobalt chrome stems while a lower incidence of stem pain was associated with slotted titanium stems. Symptoms were intermittent and activity related in all but 2 cases, but none required component revision. The presence of end-of-stem pain did appear to be clinically significant because patients with this pain were more likely to be only somewhat satisfied or dissatisfied with their degree of pain relief compared to those without end-of-stem pain. Stem design does appear to impact the incidence of end-of-stem pain in revision total knee arthroplasty.

Biomechanical study of pedicle screw fixation in severely osteoporotic bone.

BACKGROUND CONTEXT: Obtaining adequate purchase with standard pedicle screw techniques remains a challenge in poor quality bone. The development of alternate insertion techniques and screw designs was prompted by recognition of potential fixation complications. An expandable pedicle screw design has been shown to significantly improve fixation compared to a conventional screw in poor quality bone. PURPOSE: The purpose of this study was to determine if polymethylmethacrylate (PMMA) bone cement augmentation of an expandable pedicle screw can further improve fixation strength compared to the expandable screw alone in severely osteoporotic bone. A technique for cement insertion into the pedicle by means of the cannulated central portion of the expandable screw is also described. STUDY DESIGN: The axial pullout strength, stiffness and energy absorbed of cemented and noncemented expandable pedicle screws was determined in cadaveric vertebrae. METHODS: Twenty-one fresh unembalmed vertebrae from the thoracolumbar spine were used. Radiographs and bone mineral density measurements (BMD) were used to characterize bone quality. Paired cemented and noncemented pedicle screw axial pullout strength was determined through mechanical testing. Mechanical pullout strength, stiffness and energy to failure was correlated with BMD. RESULTS: Overall, there was a 250% increase in mean pullout strength with the cemented expandable screw compared with a noncemented expandable screw including a greater than twofold increase in pullout strength in the most severely osteoporotic bone. The mean stiffness and energy absorbed to failure was also significantly increased. A cemented conventional screw achieved a pullout strength similar to the noncemented expandable screw. CONCLUSIONS: PMMA cement augmentation of the expandable pedicle screw may be a viable clinical option for achieving fixation in severely osteoporotic bone.

Treating extensor mechanism disruption after total knee arthroplasty.

A consecutive series of 14 patients with chronic extensor mechanism disruption after total knee arthroplasty (TKA) were treated with allograft reconstruction. Preoperatively all patients had full passive extension but a complete extensor lag. The average time from extensor mechanism disruption to surgery was 7 months (range, 3-24 months). Two methods of reconstruction were used depending on specific indications: an Achilles tendon with calcaneal bone block (eight patients) or a quadriceps tendon-patella-patellar tendon-tibial tubercle composite graft (six patients). At followup averaging 42 months (range, 24-60 months) all patients were community ambulators, five patients used a cane, two patients used a walker, and seven patients used no assistive devices. One patient had a partial rerupture and a 45 degrees extensor lag, one patient had a 30 degrees extensor lag, two patients had a 15 degrees extensor lag, and 10 patients had a lag of less than 10 degrees. All patients thought that their functional status had been improved and were satisfied with the results of the allograft reconstruction.