Tools & Techniques - Statistics: Propensity score techniques.

Propensity score (PS) techniques are useful if the number of potential confounding pretreatment variables is large and the number of analysed outcome events is rather small so that conventional multivariable adjustment is hardly feasible. Only pretreatment characteristics should be chosen to derive PS, and only when they are probably associated with outcome. A careful visual inspection of PS will help to identify areas of no or minimal overlap, which suggests residual confounding, and trimming of the data according to the distribution of PS will help to minimise residual confounding. Standardised differences in pretreatment characteristics provide a useful check of the success of the PS technique employed. As with conventional multivariable adjustment, PS techniques cannot account for confounding variables that are not or are only imperfectly measured, and no PS technique is a substitute for an adequately designed randomised trial.

Clinical effectiveness of cognitive behavioral therapy for depression in routine care: A propensity score based comparison between randomized controlled trials and clinical practice

Background: The efficacy of cognitive behavioral therapy (CBT) for the treatment of depressive disorders has been demonstrated in many randomized controlled trials (RCTs). This study investigated whether for CBT similar effects can be expected under routine care conditions when the patients are comparable to those examined in RCTs. Method: N=574 CBT patients from an outpatient clinic were stepwise matched to the patients undergoing CBT in the National Institute of Mental Health Treatment of Depression Collaborative Research Program (TDCRP). First, the exclusion criteria of the RCT were applied to the naturalistic sample of the outpatient clinic. Second, propensity score matching (PSM) was used to adjust the remaining naturalistic sample on the basis of baseline covariate distributions. Matched samples were then compared regarding treatment effects using effect sizes, average treatment effect on the treated (ATT) and recovery rates. Results: CBT in the adjusted naturalistic subsample was as effective as in the RCT. However, treatments lasted significantly longer under routine care conditions. Limitations: The samples included only a limited amount of common predictor variables and stemmed from different countries. There might be additional covariates, which could potentially further improve the matching between the samples. Conclusions: CBT for depression in clinical practice might be equally effective as manual-based treatments in RCTs when they are applied to comparable patients. The fact that similar effects under routine conditions were reached with more sessions, however, points to the potential to optimize treatments in clinical practice with respect to their efficiency.

Long-term comparison of everolimus-eluting and sirolimus-eluting stents for coronary revascularization

Objectives This study sought to compare the unrestricted use of everolimus-eluting stents (EES) with sirolimus-eluting stents (SES) in patients undergoing percutaneous coronary intervention. Background It is unclear whether there are differences in safety and efficacy between EES and SES during long-term follow-up. Methods Using propensity score matching, clinical outcome was compared among 1,342 propensity score–matched pairs of patients treated with EES and SES. The primary outcome was a composite of death, MI, and target vessel revascularization. Results The median follow-up was 1.5 years with a maximum of 3 years. The primary outcome occurred in 14.9% of EES- and 18.0% of SES-treated patients up to 3 years (hazard ratio [HR]: 0.83, 95% confidence interval [CI]: 0.68 to 1.00, p = 0.056). All-cause mortality (6.0% vs. 6.5%, HR: 0.92, 95% CI: 0.68 to 1.25, p = 0.59) was similar, risks of myocardial infarction (MI) (3.3% vs. 5.0%, HR: 0.62, 95% CI: 0.42 to 0.92, p = 0.017), and target vessel revascularization (7.0% vs. 9.6%, HR: 0.75, 95% CI: 0.57 to 0.99, p = 0.039) were lower with EES than SES. Definite stent thrombosis (ST) (HR: 0.30, 95% CI: 0.12 to 0.75, p = 0.01) was less frequent among patients treated with EES. The reduced rate of MI with EES was explained in part by the lower risk of definite ST and the corresponding decrease in events associated with ST (HR: 0.25, 95% CI: 0.08 to 0.75, p = 0.013). Conclusions The unrestricted use of EES appears to be associated with improved clinical long-term outcome compared with SES. Differences in favor of EES are driven in part by a lower risk of MI associated with ST.

Total hip arthroplasty: leg length inequality impairs functional outcomes and patient satisfaction

Background Leg length inequality (LLI) was identified as a problem of total hip arthroplasty soon after its introduction. Leg lengthening is the most common form of LLI. Possible consequences are limping, neuronal dysfunction and aseptic component loosening. LLI can result in an increased strain both on the contralateral hip joint and on the abductor muscles. We assessed the influence of leg lengthening and shortening on walking capacity, hip pain, limping and patient satisfaction at 2-year follow-up. Methods 478 cases with postoperative lengthening and 275 with shortening were identified, and matched with three controls each. Rigorous adjustment for potential differences in baseline patient characteristics was performed by propensity-score matching of covariates. The arbitrarily defined desired outcomes were a walking capacity >60 minutes, no hip pain, no limping, and excellent patient satisfaction. Differences in not achieving the desired outcomes between the groups were expressed as odds ratios. Results In the lengthened case group, the odds ratio for not being able to walk for an hour was 1.70 (95% CI 1.28-2.26) for cases compared to controls, and the odds ratio for having hip pain at follow-up was 1.13 (95% CI 0.78-1.64). The odds ratio for limping was 2.08 (95% CI 1.55-2.80). The odds ratio for not achieving excellent patient satisfaction was 1.67 (95% CI 1.23-2.28). In the shortening case group, the odds ratio for not being able to walk for an hour was 1.23 (95% CI 0.84-1.81), and the odds ratio for having hip pain at follow-up was 1.60 (95% CI 1.05-2.44). The odds ratio for limping for cases was 2.61 (95% CI 1.78-3.21). The odds ratio for not achieving excellent patient satisfaction was 2.15 (95% CI 1.44-3.21). Conclusions Walking capacity, limping and patient satisfaction were all significantly associated with leg lengthening, whereas pain alleviation was not. In contrast, hip pain, limping and patient satisfaction were all significantly associated with leg shortening, whereas walking capacity was not.

Asthma in young south Asian women living in the United Kingdom: the importance of early life

BACKGROUND: Studies of immigrants suggest that the environment during fetal life and duration of residence in the host country might influence the development of asthma. Little is known about the importance of the timing of the exposure in the host country and whether migrants might be especially vulnerable in certain age windows. OBJECTIVE: We compared the reported prevalence of asthma between young white and south Asian women in the United Kingdom, and investigated associations with country of birth and age at immigration. METHODS: A questionnaire on atopic disorders was posted to 2380 south Asian and 5796 white young mothers randomly sampled in Leicestershire. Data on ethnicity were also available from maternity records. Data were analysed using multivariable logistic regression and a propensity score approach. Results The reported prevalence of asthma was 10.9% in south Asian and 21.8% in white women. South Asian women who migrated to the United Kingdom aged 5 years or older reported less asthma (6.5%) than those born in the United Kingdom or who migrated before age 5 (16.0%), with an adjusted odds ratio of 0.38 [95% Confidence Interval 0.23-0.64, P<0.001]. For those who migrated aged over 5 years, the prevalence did not alter with the duration of residence in the United Kingdom. Current exposure to common environmental risk factors had relatively little effect on prevalence estimates. CONCLUSION: These data from a large population-based study support the hypothesis that early life environmental factors influence the risk of adult asthma.

Minimal extracorporeal circulation is a promising technique for coronary artery bypass grafting

BACKGROUND: Minimal extracorporeal circulation (MECC) is a promising perfusion technology, taking the advantage of an ECC while having a significantly reduced priming volume. We analyzed the actual possible benefits of using MECC in patients undergoing CABG procedures and compared the results with conventional extracorporeal circulation (CECC). METHODS: One thousand fifty-three consecutive patients underwent CABG surgery using the MECC perfusion technique. Subgroup analyses focused on perioperative myocardial markers (cardiac troponin I [cTnI]), incidence of atrial fibrillation (AF), and perioperative evaluation of inflammatory markers and data were compared with those of patients who underwent CABG using CECC. A propensity score analysis was performed. RESULTS: Patient characteristics and distribution of EuroSCORE risk were similar in both groups. Severity of coronary artery disease and extent of revascularization were also comparable in both groups (number of distal anastomoses: 3.2 +/- 1.1 in CECC vs 3.2 +/- 0.9 in MECC; p = not significant [ns]). The cTnI was significantly lower in the MECC group (11.0 +/- 10.8 microg/L in MECC vs 24.7 +/- 25.3 microg/L in CECC; p < 0.05). Incidence of AF was 11.1% in MECC and 39.0% in CECC (p < 0.05). Inflammatory markers (interleukin-6, SC5b-9) were lower in MECC patients (p < 0.05). Propensity score analysis confirmed faster recovery in MECC patients and lower incidence of AF. CONCLUSIONS: Minimal extracorporeal circulation is a safe perfusion technique for CABG and may therefore concurrence OPCAB and traditional CABG under CECC.

A comparison of patient characteristics and 30-day mortality outcomes after transcatheter aortic valve implantation and surgical aortic valve replacement for the treatment of aortic stenosis: a two-centre study

AIMS: It is unclear whether transcatheter aortic valve implantation (TAVI) addresses an unmet clinical need for those currently rejected for surgical aortic valve replacement (SAVR) and whether there is a subgroup of high-risk patients benefiting more from TAVI compared to SAVR. In this two-centre, prospective cohort study, we compared baseline characteristics and 30-day mortality between TAVI and SAVR in consecutive patients undergoing invasive treatment for aortic stenosis. METHODS AND RESULTS: We pre-specified different adjustment methods to examine the effect of TAVI as compared with SAVR on overall 30-day mortality: crude univariable logistic regression analysis, multivariable analysis adjusted for baseline characteristics, analysis adjusted for propensity scores, propensity score matched analysis, and weighted analysis using the inverse probability of treatment (IPT) as weights. A total of 1,122 patients were included in the study: 114 undergoing TAVI and 1,008 patients undergoing SAVR. The crude mortality rate was greater in the TAVI group (9.6% vs. 2.3%) yielding an odds ratio [OR] of 4.57 (95%-CI 2.17-9.65). Compared to patients undergoing SAVR, patients with TAVI were older, more likely to be in NYHA class III and IV, and had a considerably higher logistic EuroSCORE and more comorbid conditions. Adjusted OR depended on the method used to control for confounding and ranged from 0.60 (0.11-3.36) to 7.57 (0.91-63.0). We examined the distribution of propensity scores and found scores to overlap sufficiently only in a narrow range. In patients with sufficient overlap of propensity scores, adjusted OR ranged from 0.35 (0.04-2.72) to 3.17 (0.31 to 31.9). In patients with insufficient overlap, we consistently found increased odds of death associated with TAVI compared with SAVR irrespective of the method used to control confounding, with adjusted OR ranging from 5.88 (0.67-51.8) to 25.7 (0.88-750). Approximately one third of patients undergoing TAVI were found to be potentially eligible for a randomised comparison of TAVI versus SAVR. CONCLUSIONS: Both measured and unmeasured confounding limit the conclusions that can be drawn from observational comparisons of TAVI versus SAVR. Our study indicates that TAVI could be associated with either substantial benefits or harms. Randomised comparisons of TAVI versus SAVR are warranted.

Naturalistic study on the effectiveness of psycho-oncological interventions in cancer patients and their partners

BACKGROUND There is evidence for the efficacy of psycho-oncological interventions (POI) in randomized controlled trials for cancer patients. Our objective was to explore, under naturalistic conditions (using propensity score matching), whether POI are effective to decrease anxiety, depression, distress and overall psychopathological symptoms within cancer patients and their partners. METHODS This study was conducted in the Oncology and Hematology Center of a University clinic in Switzerland with a group of 186 patients and 117 partners. Outcome measures of mental health were the Hospital Anxiety and Depression Scale and the Symptom Checklist (SCL-9-K). Repeated-measures ANOVAs were used to analyze change over time and group effects between individuals with POI vs. without POI. RESULTS Highly distressed patients and their partners participating in POI reported better mental health over time. Among moderately distressed patients, a decrease over time emerged in depression and distress independent of POI. No effectiveness of POI could be demonstrated in moderately distressed patients and partners. CONCLUSION Most of the highly distressed patients receive additional POI and therefore conclusions about the efficacy of POI are difficult. For moderately distressed individuals, POI as implemented in Switzerland does not improve mental health in such patients and their partners, which may be caused by very time limited POI treatments. Studies with more intense POI treatments are needed.

Comparison of pain-resilient working individuals to population-based case controls with/without momentary low back pain

Background: Few studies have examined the 20% of individuals who never experience an episode of low back pain (LBP). To date, no investigation has been undertaken that examines a group who claim to have never experienced LBP in their lifetime in comparison to two population-based case–control groups with and without momentary LBP. This study investigates whether LBP-resilient workers between 50 and 65 years had better general health, demonstrated more positive health behaviour and were better able to achieve routine activities compared with both case–control groups. Methods: Forty-two LBP-resilient participants completed the same pain assessment questionnaire as a population-based LBP sample from a nationwide, large-scale cross-sectional survey in Switzerland. The LBP-resilient participants were pairwise compared to the propensity score-matched case controls by exploring differences in demographic and work characteristics, and by calculating odds ratios (ORs) and effect sizes. A discriminant analysis explored group differences, while the multiple logistic regression analysis specified single indicators which accounted for group differences. Results: LBP-resilient participants were healthier than the case controls with momentary LBP and achieved routine activities more easily. Compared to controls without momentary LBP, LBP-resilient participants had a higher vitality, a lower workload, a healthier attitude towards health and behaved more healthily by drinking less alcohol. Conclusions: By demonstrating a difference between LBP-resilient participants and controls without momentary LBP, the question that arises is what additional knowledge can be attained. Three underlying traits seem to be relevant about LBP-resilient participants: personality, favourable work conditions and subjective attitudes/attributions towards health. These rationales have to be considered with respect to LBP prevention.

Treatment modalities for T1N0 esophageal cancers: a comparative analysis of local therapy versus surgical resection

BACKGROUND To investigate the role of nonsurgical treatment for early-stage esophageal cancer, we compared the outcomes of local therapy to esophagectomy, using a large, national database. METHODS Five-year cancer-specific and overall survival (OS) of patients, with T1N0M0 squamous cell or adenocarcinoma of the mid or distal esophagus treated with either surgery or local therapy, with ablative and/or excision techniques, in the Surveillance Epidemiology and End Results cancer registry from 1998 to 2008, were compared using the Kaplan-Meier approach, and multivariable and propensity-score adjusted Cox proportional hazard, and competing risk models. RESULTS Of 1458 patients with T1N0 esophageal cancer, 1204 (83%) had surgery and 254 (17%) had local therapy only. The use of local therapy increased significantly from 8.1% in 1998 to 24.1% in 2008 (p < 0.001). The 5-year OS after local excisional therapy and surgery was not significantly different (55.5% versus 64.1% respectively, p = 0.07), and 5-year cancer-specific survival (CSS) also did not differ (81.7% versus 75.8%, p = 0.10). However, after propensity-score adjustment, CSS was better for patients who underwent local therapy compared with those who underwent surgery (hazard ratio: 0.46, 95% confidence interval: 0.27-0.77, p = 0.003), whereas OS remained similar. CONCLUSION The use of local therapy for T1N0 esophageal cancers increased significantly from 1998 to 2008. Compared with those treated with esophagectomy, patients treated with local therapy had similar OS but improved CSS, indicating a higher chance of dying from other causes. Further studies are needed to confirm the oncologic efficacy of local therapy when used in patients whose lifespans are not limited by conditions other than esophageal cancer.

Influenza vaccination and cardiovascular risk in patients with recent TIA and stroke

OBJECTIVES: To determine whether current influenza vaccination is associated with reduced risk of major vascular events in patients with recent ischemic stroke or TIA of mainly atherothrombotic origin. METHODS: Data were pooled from 2 prospective cohort studies, the OPTIC Registry (n = 3,635) and the AMISTAD Study (n = 618), and from the randomized PERFORM Trial (n = 19,120), all of which included patients with recent ischemic stroke or TIA. Influenza vaccination status was determined in 23,110 patients. The primary outcome was a composite of nonfatal myocardial infarction, nonfatal stroke, or vascular death up to 2 years. Secondary outcomes were myocardial infarction and stroke separately. RESULTS: Influenza vaccination had no association with the primary outcome in the propensity score-matched cohort (hazard ratio 0.97, 95% confidence interval [CI] 0.85-1.11; p = 0.67) or in the propensity score-adjusted cohort (hazard ratio 1.00, 95% CI 0.89-1.12; p = 0.99). Similarly, the risk of stroke and myocardial infarction did not differ between the vaccinated group and the unvaccinated group; in the matched cohort, the hazard ratio was 1.01 (95% CI 0.88-1.17; p = 0.89) for stroke and 0.84 (95% CI 0.59-1.18; p = 0.30) for myocardial infarction. CONCLUSIONS: Influenza vaccination was not associated with reduced outcome events in patients with recent atherothrombotic ischemic stroke after considering all baseline characteristics (including concomitant medications) associated with influenza vaccination.

Long-Term Comparison of Everolimus-Eluting and Biolimus-Eluting Stents in All-Comer Patients

AIMS: Second-generation everolimus-eluting stents (EES) are safer and more efficient than first-generation paclitaxel-eluting stents (PES). Third-generation biolimus-eluting stents (BES) have been found to be non-inferior to PES. To date, there is no available comparative study between EES and BES. We aimed to investigate the safety and efficacy of BES with biodegradable polymer compared to EES with durable polymer at a follow-up of two years in an unselected population of consecutively enrolled patients. METHODS AND RESULTS: A group of 814 consecutive patients undergoing percutaneous coronary intervention (PCI) was enrolled between 2007 and 2010, of which 527 were treated with EES and 287 with BES implantation. Clinical outcome was compared in 200 pairs using propensity score matching. The primary endpoint was a composite of death, myocardial infarction (MI) and target vessel revascularisation (TVR) at two-year follow-up. Median follow-up was 22 months. The primary outcome occurred in 11.5% of EES and 10.5% of BES patients (HR 1.11, 95% CI: 0.61-2.00, p=0.74). At two years, there was no significant difference with regard to death (HR 0.49, 95% CI: 0.18-1.34, p=0.17), cardiac death (HR 0.14, 95% CI: 0.02-1.14, p=0.66) or MI (HR 6.10, 95% CI: 0.73-50.9, p=0.10). Stent thrombosis (ST) incidence was evenly distributed between EES (n=2) and BES (n=2) (p-value=1.0). CONCLUSIONS: This first clinical study failed to demonstrate any significant difference regarding safety or efficacy between these two types and generations of drug-eluting stents (DES).

Drug-eluting stents compared to bare-metal stents improve short-term survival in patients with acute myocardial infarction undergoing primary percutaneous coronary intervention: a nationwide prospective analysis of the AMIS Plus registry

BACKGROUND Recently, it has been suggested that the type of stent used in primary percutaneous coronary interventions (pPCI) might impact upon the outcomes of patients with acute myocardial infarction (AMI). Indeed, drug-eluting stents (DES) reduce neointimal hyperplasia compared to bare-metal stents (BMS). Moreover, the later generation DES, due to its biocompatible polymer coatings and stent design, allows for greater deliverability, improved endothelial healing and therefore less restenosis and thrombus generation. However, data on the safety and performance of DES in large cohorts of AMI is still limited. AIM To compare the early outcome of DES vs. BMS in AMI patients. METHODS This was a prospective, multicentre analysis containing patients from 64 hospitals in Switzerland with AMI undergoing pPCI between 2005 and 2013. The primary endpoint was in-hospital all-cause death, whereas the secondary endpoint included a composite measure of major adverse cardiac and cerebrovascular events (MACCE) of death, reinfarction, and cerebrovascular event. RESULTS Of 20,464 patients with a primary diagnosis of AMI and enrolled to the AMIS Plus registry, 15,026 were referred for pPCI and 13,442 received stent implantation. 10,094 patients were implanted with DES and 2,260 with BMS. The overall in-hospital mortality was significantly lower in patients with DES compared to those with BMS implantation (2.6% vs. 7.1%,p < 0.001). The overall in-hospital MACCE after DES was similarly lower compared to BMS (3.5% vs. 7.6%, p < 0.001). After adjusting for all confounding covariables, DES remained an independent predictor for lower in-hospital mortality (OR 0.51,95% CI 0.40-0.67, p < 0.001). Since groups differed as regards to baseline characteristics and pharmacological treatment, we performed a propensity score matching (PSM) to limit potential biases. Even after the PSM, DES implantation remained independently associated with a reduced risk of in-hospital mortality (adjusted OR 0.54, 95% CI 0.39-0.76, p < 0.001). CONCLUSIONS In unselected patients from a nationwide, real-world cohort, we found DES, compared to BMS, was associated with lower in-hospital mortality and MACCE. The identification of optimal treatment strategies of patients with AMI needs further randomised evaluation; however, our findings suggest a potential benefit with DES.

Survival effects of inferior vena cava filter in patients with acute symptomatic venous thromboembolism and a significant bleeding risk.

OBJECTIVES The purpose of this study was to investigate the survival effects of inferior vena cava filters in patients with venous thromboembolism (VTE) who had a significant bleeding risk. BACKGROUND The effectiveness of inferior vena cava filter use among patients with acute symptomatic VTE and known significant bleeding risk remains unclear. METHODS In this prospective cohort study of patients with acute VTE identified from the RIETE (Computerized Registry of Patients With Venous Thromboembolism), we assessed the association between inferior vena cava filter insertion for known significant bleeding risk and the outcomes of all-cause mortality, pulmonary embolism (PE)-related mortality, and VTE rates through 30 days after the initiation of VTE treatment. Propensity score matching was used to adjust for the likelihood of receiving a filter. RESULTS Of the 40,142 eligible patients who had acute symptomatic VTE, 371 underwent filter placement because of known significant bleeding risk. A total of 344 patients treated with a filter were matched with 344 patients treated without a filter. Propensity score-matched pairs showed a nonsignificant trend toward lower risk of all-cause death for filter insertion compared with no insertion (6.6% vs. 10.2%; p = 0.12). The risk-adjusted PE-related mortality rate was lower for filter insertion than no insertion (1.7% vs. 4.9%; p = 0.03). Risk-adjusted recurrent VTE rates were higher for filter insertion than for no insertion (6.1% vs. 0.6%; p < 0.001). CONCLUSIONS In patients presenting with VTE and with a significant bleeding risk, inferior vena cava filter insertion compared with anticoagulant therapy was associated with a lower risk of PE-related death and a higher risk of recurrent VTE. However, study design limitations do not imply a causal relationship between filter insertion and outcome.

Scarring Effects of Early Unemployment among Young Workers with Vocational Credentials in Switzerland

Background Using a unique, longitudinal survey that follows school-to-work transitions of pupils who participated in PISA 2000, this paper investigates adverse consequences, so-called scarring effects, of early unemployment among young adults who acquired vocational credentials in Switzerland. Methods As social, individual and contextual factors influence both early unemployment and later employment outcomes, taking into account endogeneity is of utmost importance when investigating scarring effects. In this regard we make use of nearest-neighbour propensity score matching and set up statistical control groups. Results Our results suggest that young adults who hold vocational credentials are more likely to be neither in employment nor in education, and to earn less and be more dissatisfied with their career progress later in work life than they would be, had they not experienced early unemployment. Conclusions We conclude that unemployment scarring also affects young adults with vocational credentials in a liberal labour market setting that otherwise allows for smooth school-to-work transitions. This finding runs counter to expectations that standardised vocational degrees, a liberal and flexible labour market structure, and predominantly short unemployment spells protect young skilled workers from scarring in case they happen to experience early career instability.