Absolute Standardized Mean Differences Research Articles

Optimization-Based Stable Balancing Weights Versus Propensity Score Weighting for Samples With High Covariate Imbalance.

To compare the performance (covariate balance, effective sample size [ESS]) of stable balancing weights (SBW) versus propensity score weighting (PSW). Two applied cases were used to compare performance: (Case 1) extreme imbalance in baseline covariates between groups and (Case 2) substantial discrepancy in sample size between groups. Using the Premier Healthcare Database, we selected patients who (Case 1) underwent a surgical procedure with one of two different bipolar forceps between January 2000 and June 2020, or (Case 2) a neurological procedure using one of two different nonabsorbable surgical sutures between January 2000 and March 2020. Average treatment effects on the treated (ATT) weights were generated based on selected covariates. SBW was implemented using two techniques: (1) "grid search" to find weights of minimum variance at the lowest target absolute standardized mean difference (SMD); (2) finding weights of minimum variance at prespecified SMD tolerance. PSW and SBW methods were compared on postweighting SMDs, the number of imbalanced covariates, and ESS of the ATT-weighted control group. In both studies, improved covariate balance was achieved with both SBW techniques. All methods suffered from postweighting ESS that was lower than the unweighted control group's original sample size; however, SBW methods achieved higher ESS for the control groups. Sensitivity analyses using SBW to apply variable-specific SMD thresholds increased ESS, outperforming PSW. In this applied example, the optimization-based SBW method provided ample flexibility with respect to prespecification of covariate balance goals and resulted in better postweighting covariate balance and larger ESS as compared with PSW.

Characteristics of Patients Hospitalized in Rural and Urban ICUs From 2010 to 2019.

Rural hospitals are threatened by workforce shortages and financial strain. To optimize regional critical care delivery, it is essential to understand what types of patients receive intensive care in rural and urban hospitals. A retrospective cohort study. All fee-for-service Medicare beneficiaries in the United States who were 65 years old or older hospitalized in an ICU between 2010 and 2019 were included. Rural and urban hospitals were classified according to the 2013 National Center For Health Statistics Urban-Rural Classification Scheme for Counties. Patient comorbidities, primary diagnoses, organ dysfunction, and procedures were measured using the International Classification of Diseases, 9th and 10th revisions diagnosis and procedure codes. Standardized differences were used to compare rural and urban patient admission characteristics. None. There were 12,224,097 ICU admissions between 2010 and 2019, and 1,488,347 admissions (12.2%) were to rural hospitals. The most common diagnoses in rural hospitals were cardiac (30.3%), infectious (24.6%), and respiratory (10.9%). Patients in rural ICUs had similar organ dysfunction compared with urban hospitals (mean organ failures in rural ICUs 0.5, sd 0.8; mean organ failures in urban ICUs 0.6, sd 0.9, absolute standardized mean difference 0.096). Organ dysfunction among rural ICU admissions increased over time (0.4 mean organ failures in 2010 to 0.7 in 2019, p < 0.001). Rural hospitals care for an increasingly complex critically ill patient population with similar organ dysfunction as urban hospitals. There is a pressing need to develop policies at federal and regional healthcare system levels to support the continued provision of high-quality ICU care within rural hospitals.

Comparative effectiveness of linvoseltamab versus standard-of-care (SOC) treatment (tx) in real-world patients (pts) in the United States (US) with triple-class exposed (TCE) relapsed/refractory multiple myeloma (RRMM).

7561 Background: Pts with TCE RRMM have poor outcomes and a high unmet need, with no established SOC tx. LINKER-MM1 (NCT03761108) is a single-arm, Phase 1/2 study investigating linvoseltamab, a B-cell maturation antigen × CD3 bispecific antibody, in pts with RRMM who were previously treated with a proteasome inhibitor, immunomodulatory drug, and anti-CD38 antibody, or were triple-class refractory (TCR) to these tx. The aim of this study was to contextualize LINKER-MM1 by comparing outcomes with linvoseltamab vs a real-world (RW) external control arm (ECA). Methods: A RW ECA was derived from 2 US electronic health record databases (COTA, Guardian Research Network) of pts who started a new line of therapy (LOT) after classification as TCE or TCR and met key eligibility criteria for LINKER-MM1. Eligibility was assessed at the initiation of each new LOT, and all eligible LOTs were included in the ECA. Data in Phase 2 pts who received linvoseltamab 200 mg in LINKER-MM1 were included (data cutoff: Sep 8, 2023). Inverse probability of tx weighting (IPTW) was used to reduce imbalances between the RW and LINKER-MM1 cohorts. Key prognostic factors were identified using a systematic review and rank ordered by an international committee of MM experts (Kumar et al., 2023). Outcomes included overall response rate (ORR), progression-free survival (PFS), time to next treatment (TTNT), and overall survival (OS). An independent committee of epidemiology and oncology experts reviewed the comparability of the cohorts and endpoint assessments prior to conducting comparative analyses. Results: Comparative analyses were performed in 105 pts in the linvoseltamab cohort and 101 RW pts (137 LOTs). Following IPTW, the distribution of cytogenetic risk, age, TCR status, Eastern Cooperative Oncology Group Performance Status, and platelet count were balanced between the 2 cohorts (absolute standardized mean difference &lt;0.10). After adjustment, pts receiving linvoseltamab had significantly improved ORR, PFS, TTNT, and OS vs the RW ECA (see table). Conclusions: Linvoseltamab significantly improved outcomes vs RW SOC in the US, highlighting its potential as a highly effective tx in pts with TCE RRMM. [Table: see text]

A simulation-based bias analysis to assess the impact of unmeasured confounding when designing non-randomized database studies.

Unmeasured confounding is often raised as a source of potential bias during the design of non-randomized studies but quantifying such concerns is challenging. We developed a simulation-based approach to assess the potential impact of unmeasured confounding during the study design stage. The approach involved generation of hypothetical individual-level cohorts using realistic parameters including a binary treatment (prevalence 25%), a time-to-event outcome (incidence 5%), 13 measured covariates, a binary unmeasured confounder (u1, 10%), and a binary measured 'proxy' variable (p1) correlated with u1. Strength of unmeasured confounding and correlations between u1 and p1 were varied in simulation scenarios. Treatment effects were estimated with, a) no adjustment, b) adjustment for measured confounders (Level 1), c) adjustment for measured confounders and their proxy (Level 2). We computed absolute standardized mean differences in u1 and p1 and relative bias with each level of adjustment. Across all scenarios, Level 2 adjustment led to improvement in balance of u1, but this improvement was highly dependent on the correlation between u1 and p1. Level 2 adjustments also had lower relative bias than Level 1 adjustments (in strong u1 scenarios: relative bias of 9.2%, 12.2%, 13.5% at correlations 0.7, 0.5, and 0.3, respectively versus 16.4%, 15.8%, 15.0% for Level 1, respectively). An approach using simulated individual-level data was useful to explicitly convey the potential for bias due to unmeasured confounding while designing non-randomized studies and can be helpful in informing design choices.

#462 Initiation of SGLT2i vs GLP1-RA and incidence of AKI in persons with type 2 diabetes mellitus

Abstract Background and Aims Adverse kidney outcomes, including acute kidney injury (AKI), are frequent complications in people with diabetes mellitus type 2. In large placebo-controlled randomized trials, both sodium-glucose cotransporter-2 inhibitors (SGLT2i) and glucagon-like peptide-1 receptor agonists (GLP1-RA) have shown efficacy in reducing the risk of kidney outcomes; however, no trial has compared the kidney protective effects of these drug classes head-to-head. We therefore conducted an emulated target trial to compare the effect of SGLT2i to that of GLP1-RA on the incidence of AKI. Method Using routinely collected data from nationwide Danish population-based registries, we included all adults with diabetes mellitus type 2 initiating SGLT2i or GLP1-RA in 2015-2020. Patients with prevalent kidney failure (defined as eGFR &amp;lt; 15 mL/min/1.73m2, dialysis, or kidney transplant), and patients initiating GLP1-RA marketed specifically for treatment of obesity were excluded. We assessed demographic and diabetes-related baseline characteristics along with comorbidities and comedication. We used propensity scores, estimated by logistic regression, to compute stabilized inverse probability of treatment weights, which were applied in all analyses to reduce confounding. Using an intention to treat approach, we followed patients from treatment initiation until the first of death, emigration, or end of follow-up (31 December 2021). Using creatinine measurements from in- and outpatient settings, we applied the KDIGO definition of AKI to record each episode of AKI during follow-up. We estimated the risk of a first AKI episode over time using the Aalen-Johansen estimator. Furthermore, to estimate the mean number of AKIs over time, i.e., including recurrent AKIs, we used the mean cumulative count estimator. Both estimators account for the competing risk of death. Comparing SGLT2i to GLP1-RA, we reported differences in risks and counts at 1, 3 and 5 years of follow-up. Finally, we estimated all-cause mortality using the Kaplan-Meier estimator. We used bootstrapping to obtain 95% confidence intervals (95% CI). Results We included 68 010 individuals, of which 44 034 (65%) initiated SGLT2i, and 23 976 (35%) initiated GLP1-RA. Initiators of SGLT2i were slightly older (median age 63 vs 62 years) and more often male (63% vs 56%) than initiators of GLP1-RA, but with similar diabetes duration (7 vs 7 years). The annual number of initiators of each drug in Denmark increased over the study period. Compared to GLP1-RA, initiators of SGLT2i generally had better estimated glomerular filtration rate (eGFR) at baseline (9% vs 18% had eGFR &amp;lt; 60 mL/min/1.73m2), more frequently used metformin (91% vs 84%), but less frequently insulin (11% vs 28%). Initiators of SGLT2i had a lower prevalence of hospital-diagnosed retinopathy (7% vs 9%), neuropathy (5% vs 6%), and obesity (13% vs 23%) than initiators of GLP1-RA. After inverse probability of treatment weighting, each variable was well balanced with absolute standardized mean differences below 0.02. During a median follow-up of 2.8 years, the weighted absolute risks of first time AKI at 1, 3, and 5 years were 4.8%, 11.2%, and 17.0% for initiators of SGLT2i, and 5.4%, 12.3%, and 19.1% for initiators of GLP1-RA, resulting in AKI risk differences of −0.6% (95% CI: −1.0% to −0.3%), −1.0% (95% CI: −1.7% to −0.5%), and −2.0% (95% CI: −3.0% to −1.1%). The mean cumulative counts showed a similar pattern, with SGLT2i initiators experiencing slightly fewer AKI episodes than GLP1-RA initiators on average (Fig. 1). After 5 years, an estimated 10.3% of SGLT2i initiators and 9.6% of GLP1-RA initiators had died, resulting in a risk difference of 0.7% (95% CI: -0.1% to 1.5%). Conclusion People with diabetes mellitus type 2 who initiated SGLT2i had slightly lower weighted incidence of AKI over 5 years than those initiating GLP1-RA.

Postoperative outcomes and costs of laparoscopic versus robotic distal pancreatectomy: a propensity-matched analysis.

Minimally invasive distal pancreatectomy (MIDP) has established advantages over the open approach. The costs associated with robotic DP (RDP) versus laparoscopic DP (LDP) make the robotic approach controversial. We sought to compare outcomes and cost of LDP and RDP using propensity matching analysis at our institution. Patients undergoing LDP or RDP between 2000 and 2021 were retrospectively identified. Patients were optimally matched using age, gender, American Society of Anesthesiologists status, body mass index, and tumor size.Between-group differences were analyzed using the Wilcoxon signed-rank test for continuous data, and the McNemar's test for categorical data. Outcomes included operative duration, conversion to open surgery, postoperative length of stay, pancreatic fistula rate, pseudocyst requiring intervention, and costs. 298 patients underwent MIDP, 180 (60%) were laparoscopic and 118 (40%) were robotic. All RDPs were matched 1:1 to a laparoscopic case with absolute standardized mean differences for all matching covariates below 0.10, except for tumor type (0.16). RDP had longer operative times (268 vs 178min, p < 0.01), shorter length of stay (2 vs 4days, p < 0.01), fewer biochemical pancreatic leaks (11.9% vs 34.7%, p < 0.01), and fewer interventional radiological drainage (0% vs 5.9%, p = 0.01). The number of pancreatic fistulas (11.9% vs 5.1%, p = 0.12), collections requiring antibiotics or intervention (11.9% vs 5.1%, p = 0.12), and conversion rates (3.4% vs 5.1%, p = 0.72) were comparable between the two groups. The total direct index admission costs for RDP were 1.01 times higher than for LDP for FY16-19 (p = 0.372), and 1.33 times higher for FY20-22 (p = 0.031). Although RDP required longer operative times than LDP, postoperative stays were shorter. The procedure cost of RDP was modestly more expensive than LDP, though this was partially offset by reduced hospital stay and reintervention rate.

Prevalence, severity, and risk factors of cancer-related fatigue among working cancer survivors: a systematic review and meta-analysis.

The aim of this study is to evaluate the prevalence, severity, and risk factors of cancer-related fatigue among currently working cancer survivors. We searched the PubMed, Embase, Scopus, CINAHL, Cochrane Library, and ICHUSHI databases. The risk of bias was evaluated independently using the Risk of Bias Assessment Tool for Non-randomized Studies (RoBANS). A meta-analysis was conducted to determine the prevalence, severity, and related factors associated with cancer-related fatigue among currently working cancer survivors. Our meta-analysis included 18 studies and revealed that 42.2% of currently working cancer survivors experience cancer-related fatigue. The fatigue severity in this group was significantly higher than that in workers without cancer (absolute standardized mean difference (SMD) = 0.67), but lower than that in cancer survivors who had previously worked and were not currently working (absolute SMD = 0.72). Distress was identified as a potential risk factor for cancer-related fatigue in working cancer survivors (partial correlation coefficient = 0.38). The high prevalence of cancer-related fatigue among employed cancer survivors underscores the need for targeted workplace interventions and fatigue management strategies. While the severity of fatigue is less than that seen in non-working survivors, the comparison with the general working population highlights a significant health disparity. The association between distress and fatigue suggests the necessity for a holistic approach to fatigue management that considers both physical and mental factors in working cancer survivors. Our findings highlight the critical need for healthcare professionals and employers to monitor fatigue levels among working cancer survivors and offer appropriate support.

Impact on Inpatient Length of Stay in Adults with Deep Partial-Thickness Burns: Comparing the Bioengineered Allogeneic Cellularized Construct Expanded-Access Trial with National Burn Repository Data.

To investigate the effect of StrataGraft (bioengineered allogeneic cellularized construct [BACC]) treatment on inpatient length of stay (LOS) as an indicator of hospital resource utilization. Data from the single-arm StrataCAT trial for adult patients with deep partial-thickness (DPT) burns who received BACC were compared with data from a matched external control arm comprising patients who received autografting for burn treatment from the National Burn Repository (NBR) during the same time period as StrataCAT. A matching, quasi-experimental approach was used to investigate the cause-and-effect relationship between BACC treatment and LOS (days). Matching factors included sex, age, ethnicity, race, burn causes, %TBSA burned (third-degree), %TBSA burned (second- and third-degrees), inhalation injury, diabetes mellitus, and hypertension. Balance was assessed between the cohorts for each confounder by standardized mean differences (SMD). Outcome was reported as average treatment effect on the treated. The BACC and NBR Autograft cohorts included 47 and 2641 patients, respectively. Following matching, the Autograft cohort had 137 patients and was weighted to 47 patients. Patients in the BACC and final (matched) Autograft cohorts were similar in all demographic and clinical covariate categories after matching (ie, the absolute SMD were < 0.1). Treatment with BACC reduced the inpatient LOS by an average of 4.84 days (P = 0.0127) relative to the comparable (matched) Autograft cohort. An ad hoc analysis revealed that mean [SD] LOS for BACC and the weighted Autograft cohorts were 17.68 [12.75] and 22.51 [19.75] days, respectively, and were 1.39 [0.94] and 1.88 [1.31] days per %TBSA burned, respectively. The significantly reduced inpatient LOS observed with BACC compared to Autograft in adults with DPT burns may translate into reduced burden on the healthcare system, reduced costs for inpatient burn treatment, and clinical benefits for patients.

The impact of COVID‐19 lockdown on mortality rate in memory clinic patients

AbstractBackgroundPeople with (pre‐)dementia were particularly at risk for mortality during the COVID‐19 pandemic due to severe outcome of a COVID‐19 infection and the negative consequences of restrictive measures. Here we study whether patients during the COVID‐19 pandemic had a higher mortality rate compared to historical controls.MethodIn this case‐control study, we included patients from the Amsterdam Dementia Cohort: 1) Pandemic patients (n = 1022) with a baseline visit between January 2017 and December 2018 and 2) Historical controls (n = 930) with a baseline visit between January 2015 and December 2016. Both groups were followed for three years (pandemic patient until 2021 and historical controls until 2019). Absolute standardized mean differences were analyzed in order to determine whether groups were well‐balanced in terms of demographics and cognition at baseline. Cox proportional hazards analyses were performed to compare mortality rate during COVID‐19 pandemic with historical controls. Age, sex, diagnosis, MMSE score, education level, and comorbidity were added as covariates to the model.ResultBoth groups were well‐balanced, as mean age of the pandemic patients was 63±9 years, n = 423 (41%) were female and mean MMSE score was 24.5±5.0. Mean age of the historical controls was 63±8 years, n = 404 (43%) were female and MMSE score was 24.2±5.2. The absolute standardized mean differences were under 0.1 for all selected covariates. Among the pandemic patients, n = 180 (18%) patients died during the previously described follow‐up time. In the historical control group, n = 131 (14%) patients died. Cox proportional hazards models showed that pandemic patients had higher risk for mortality than historical controls (HR [95%‐CI] = 1.27 [1.02‐1.59]). Results remained significant after adjusting for covariates age, sex, diagnosis, MMSE score, education level, and comorbidity (HR [95%‐CI] = 1.37 [1.09‐1.72]).ConclusionMemory clinic patients had increased risk on mortality during times of COVID‐19 lockdown compared to historical controls. This indicates that COVID‐19 lockdown regulations have accelerated disease. Even though the cause of mortality is unknown, it is likely that the excess mortality is not only due to COVID‐19 infections, but also caused by the restrictive measure of the COVID‐19 lockdown restrictions (e.g. social isolation and loss of structure).

Colchicine for cardiovacular risk reduction in lower extremity peripheral artery disease

Abstract Introduction Although recent evidence from randomized trials suggests colchicine can reduce major adverse cardiovascular events (MACE) in patients with coronary artery disease, its effect in lower extremity peripheral artery disease (PAD) is not known.1,2 To make inferences about the effects of colchicine in PAD, we leveraged the variable prescribing practice of adding colchicine versus a non-steroidal anti-inflammatory (NSAID) to urate lowering therapy in patients with gout and PAD. Purpose Estimate the effect of colchicine therapy initiation on adverse cardiovascular and limb events in a real-world sample of high-risk patients with PAD and gout. Methods A pragmatic open-label target trial comparing colchicine initiators (treatment group) to NSAID initiators (control group) was emulated in PAD patients taking either drug in conjunction with initiation of urate lowering therapy. Patients were enrolled between July 2007 and December 2019 using electronic health records from a multi-hospital academic medical system linked to Medicare claims data. Patients with contraindications to colchicine or NSAIDs (e.g., severe kidney disease, hepatic failure) were excluded. Inverse probability of treatment weighting was used to create a pseudopopulation balanced with respect to risk factors for the outcomes. The primary outcome was a 5-year composite of major adverse limb events (MALE), MACE, and all-cause mortality. Secondary outcomes included MALE and death, MACE and death, and individual components of the primary outcome accounting for the competing event of death. Risks were estimated with the Kaplan-Meier estimator for the composite outcomes and the Aalen-Johansen estimator for the competing event outcomes. Percentile based 95% confidence intervals were generated using nonparametric bootstrapping. Results A total of 1,786 patients were enrolled; mean age was 77 years (SD 7), 33% were female, and 9% were non-white race (Table). Maximum absolute standardized mean difference in relevant covariates was 23% prior to weighting and 0.8% after weighting, indicating covariate balance. The risk of the primary composite outcome of MALE, MACE, and death at 5 years was 58% (95% CI 55%, 61%) in the colchicine group and 60% (95% CI 56%, 64%) in the control group, with a risk difference of -2% (-8%, 4%) and risk ratio of 0.96 (0.89, 1.04; Figure). Similar findings were noted in the secondary composite outcomes (Figure). In the analysis accounting for the competing risk of death, the cause-specific risk of MALE (9% colchicine vs. 9% control) and MACE (24% colchicine vs. 26% control) alone were similar at 5 years. Conclusions In a real-world sample of high-risk patients with PAD and gout, colchicine did not appear to reduce the risk of adverse cardiovascular or limb events. The cardiovascular benefits of colchicine in older populations with PAD and advanced systematic atherosclerosis remain uncertain.Table.Patient characteristics.Figure.Weighted Kaplan-Meier estimates.

Factors associated with COVID-19 vaccination during June-October 2021: A multi-site prospective study.

Vaccine hesitancy presents a challenge to COVID-19 control efforts. To identify beliefs associated with delayed vaccine uptake, we developed and implemented a vaccine hesitancy survey for the COVID-19 Community Research Partnership. In June 2021, we assessed attitudes and beliefs associated with COVID-19 vaccination using an online survey. Self-reported vaccination data were requested daily through October 2021. We compared responses between vaccinated and unvaccinated respondents using absolute standardized mean differences (ASMD). We assessed validity and reliability using exploratory factor analysis and identified latent factors associated with a subset of survey items. Cox proportional hazards models and mediation analyses assessed predictors of subsequent vaccination among those initially unvaccinated. In June 2021, 29,522 vaccinated and 1,272 unvaccinated participants completed surveys. Among those unvaccinated in June 2021, 559 (43.9%) became vaccinated by October 31, 2021. In June, unvaccinated participants were less likely to feel "very concerned" about getting COVID-19 than vaccinated participants (10.6% vs. 43.3%, ASMD 0.792). Among those initially unvaccinated, greater intent to become vaccinated was associated with getting vaccinated and shorter time to vaccination. However, even among participants who reported no intention to become vaccinated, 28.5% reported vaccination before study end. Two latent factors predicted subsequent vaccination-being 'more receptive' was derived from motivation to protect one's own or others' health and resume usual activities; being 'less receptive' was derived from concerns about COVID-19 vaccines. In a Cox model, both factors were partially mediated by vaccination intention. This study characterizes vaccine hesitant individuals and identifies predictors of eventual COVID-19 vaccination through October 31, 2021. Even individuals with no intention to be vaccinated can shift to vaccine uptake. Our data suggest factors of perceived severity of COVID-19 disease, vaccine safety, and trust in the vaccine development process are predictive of vaccination and may be important opportunities for ongoing interventions.

Representativeness of the American Spine Registry: a comparison of patient characteristics with the National Inpatient Sample.

The American Spine Registry (ASR) is a collaborative effort between the American Academy of Orthopaedic Surgeons and the American Association of Neurological Surgeons. The goal of this study was to evaluate how representative the ASR is of the national practice with spinal procedures, as recorded in the National Inpatient Sample (NIS). The authors queried the NIS and the ASR for cervical and lumbar arthrodesis cases performed during 2017-2019. International Classification of Diseases, 10th Revision and Current Procedural Terminology codes were used to identify patients undergoing cervical and lumbar procedures. The two groups were compared for the overall proportion of cervical and lumbar procedures, age distribution, sex, surgical approach features, race, and hospital volume. Outcomes available in the ASR, such as patient-reported outcomes and reoperations, were not analyzed due to nonavailability in the NIS. The representativeness of the ASR compared to the NIS was assessed via Cohen's d effect sizes, and absolute standardized mean differences (SMDs) of < 0.2 were considered trivial, whereas > 0.5 were considered moderately large. A total of 24,800 arthrodesis procedures were identified in the ASR for the period between January 1, 2017, and December 31, 2019. During the same time period, 1,305,360 cases were recorded in the NIS. Cervical fusions comprised 35.9% of the ASR cohort (8911 cases) and 36.0% of the NIS cohort (469,287 cases). The two databases presented trivial differences in terms of patient age and sex for all years of interest across both cervical and lumbar arthrodeses (SMD < 0.2). Trivial differences were also noted in the distribution of open versus percutaneous procedures of the cervical and lumbar spine (SMD < 0.2). Among lumbar cases, anterior approaches were more common in the ASR than in the NIS (32.1% vs 22.3%, SMD = 0.22), but the discrepancy among cervical cases in the two databases was trivial (SMD = 0.03). Small differences were illustrated in terms of race, with SMDs < 0.5, and a more significant discrepancy was identified in the geographic distribution of participating sites (SMDs of 0.7 and 0.74 for cervical and lumbar cases, respectively). For both of these measures, SMDs in 2019 were smaller than those in 2018 and 2017. The ASR and NIS databases presented a very high similarity in proportions of cervical and lumbar spine surgeries, as well as similar distributions of age and sex, and distribution of open versus endoscopic approach. Slight discrepancies in anterior versus posterior approach among lumbar cases and patient race, and more significant discrepancies in geographic representation were also identified, yet decreasing trends in differences suggested the improving representativeness of the ASR over the course of time and its progressive growth. These conclusions are important to underline the external validity of quality investigations and research conclusions to be drawn from analyses in which the ASR is used.

Medications for alcohol use disorder improve survival in patients with hazardous drinking and alcohol-associated cirrhosis.

Medications for alcohol use disorder (MAUD) are highly effective in achieving and maintaining abstinence in patients with alcohol use disorder (AUD). Our aim was to evaluate the effect of MAUD on all-cause mortality in patients with alcohol-associated cirrhosis and active alcohol use. This was a retrospective cohort study of patients with alcohol-associated cirrhosis and high-risk alcohol use disorder in the Veterans Outcomes and Costs Associated with Liver Disease (VOCAL) database. Propensity score matching for exposure to MAUD (acamprosate or naltrexone) within a year after cirrhosis diagnosis was performed to account for potential confounders, and the association between MAUD and all-cause mortality was subsequently evaluated using Cox regression analysis. A total of 9131 patients were included, of whom 886 (9.7%) were exposed to MAUD (naltrexone: 520, acamprosate: 307, both medications: 59). The duration of MAUD exposure was >3 months in 345 patients (39%). The strongest positive predictor of MAUD prescription was an inpatient diagnosis code for AUD, followed by a concurrent diagnosis of depression; the strongest negative predictor was a history of cirrhosis decompensation. After propensity score matching (866 patients in each group) with excellent covariate balance (absolute standardized mean differences <0.1), MAUD exposure was associated with improved survival, with an HR of 0.80 relative to no MAUD exposure (95% CI: 0.67-0.97, p = 0.024). MAUD are underutilized in patients with alcohol-associated cirrhosis with high-risk alcohol use behavior but are associated with improved survival after adjustment for confounders such as the severity of liver disease, age, and engagement in the healthcare system.

Comparison of Spleen-Preservation vs Splenectomy in Minimally Invasive Distal Pancreatectomy: A Propensity-Matched Analysis

Introduction: Spleen-preservation during minimally invasive distal pancreatectomy (MIDP) can be technically challenging and remains controversial. Our primary aim was to compare minimally invasive distal pancreatectomy and splenectomy (MDPS) with minimally invasive spleen-preserving distal pancreatectomy (MSPDP). Secondary, we compared 2 spleen-preserving techniques. Methods: Adults undergoing MSPDP or MDPS (2001-2021) were retrospectively included in this single-center study. Patients were optimally matched using age, gender, American Society of Anesthesiologists score, BMI, and tumor size using MatchIt. Intraoperative and postoperative outcomes were compared using the Mann-Whitney U test for continuous data, and Fisher’s exact test for categorical data. Results: Of the 296 patients who underwent MIDP, 173 (58%) had spleen-preservation via the Warshaw technique (129;75%) or vessel-preservation (44;25%). All MDPSs were matched 1:1 to a spleen-preserving case with absolute standardized mean differences for all matching covariates below 0.20. The MSPDP group had shorter length of stay (3 vs 4 days, p = 0.01), fewer conversions to open (1 vs 13, p < 0.01) and less blood loss (p < 0.01) compared with MDPS. Operative times were comparable (245 vs 214 minutes, p = 0.16). No differences were found between the two spleen-preserving techniques. At a median follow-up of 42 (interquartile range 18-19) months, only 2 (1.2%) patients after MSPDP had required splenectomy (1 Warshaw for abscess, 1 vessel-preserving for variceal bleeding). Conclusion: Spleen-preservation does not carry the price of greater risk of blood loss, longer hospital stay, conversion, nor lengthy operating room times. Late splenectomy is very rarely required. Given the immune consequences of splenectomy, spleen-preservation should be preferred in MIDP when oncologically appropriate.

Top Surgery and Chest Dysphoria Among Transmasculine and Nonbinary Adolescents and Young Adults

Transgender and nonbinary (TGNB) adolescents and young adults (AYA) designated female at birth (DFAB) experience chest dysphoria, which is associated with depression and anxiety. Top surgery may be performed to treat chest dysphoria. To determine whether top surgery improves chest dysphoria, gender congruence, and body image in TGNB DFAB AYA. This is a nonrandomized prospective cohort study of patients who underwent top surgery between December 2019 and April 2021 and a matched control group who did not receive surgery. Patients completed outcomes measures preoperatively and 3 months postoperatively. This study took place across 3 institutions in a single, large metropolitan city. Patients aged 13 to 24 years who presented for gender-affirming top surgery were recruited into the treatment arm. Patients in the treatment arm were matched with individuals in the control arm based on age and duration of testosterone therapy. Patients in the surgical cohort underwent gender-affirming mastectomy; surgical technique was at the discretion of the surgeon. Patient-reported outcomes were collected at enrollment and 3 months postoperatively or 3 months postbaseline for the control cohort. The primary outcome was the Chest Dysphoria Measure (CDM). Secondary outcomes included the Transgender Congruence Scale (TCS) and Body Image Scale (BIS). Baseline demographic and surgical variables were collected, and descriptive statistics were calculated. Inverse probability of treatment weighting (IPTW) was used to estimate the association of top surgery with outcomes. Probability of treatment was estimated using gradient-boosted machines with the following covariates: baseline outcome score, age, gender identity, race, ethnicity, insurance type, body mass index, testosterone use duration, chest binding, and parental support. Overall, 81 patients were enrolled (mean [SD] age, 18.6 [2.7] years); 11 were lost to follow-up. Thirty-six surgical patients and 34 matched control patients completed the outcomes measures. Weighted absolute standardized mean differences were acceptable between groups with respect to body mass index, but were not comparable with respect to the remaining demographic variables baseline outcome measures. Surgical complications were minimal. IPTW analyses suggest an association between surgery and substantial improvements in CDM (-25.58 points; 95% CI, -29.18 to -21.98), TCS (7.78 points; 95% CI, 6.06-9.50), and BIS (-7.20 points; 95% CI, -11.68 to -2.72) scores. Top surgery in TGNB DFAB AYA is associated with low complication rates. Top surgery is associated with improved chest dysphoria, gender congruence, and body image satisfaction in this age group.

Comparative Adherence and Persistence of Single- and Multiple-Inhaler Triple Therapies Among Patients with Chronic Obstructive Pulmonary Disease in an English Real-World Primary Care Setting.

PurposeTriple therapy comprising a long-acting muscarinic antagonist, long-acting β2-agonist and inhaled corticosteroid is recommended for patients with chronic obstructive pulmonary disease (COPD) who continue to experience frequent exacerbations or symptoms whilst receiving dual therapy. Adherence and persistence to multiple-inhaler triple therapy (MITT) is known to be poor. This study assessed comparative adherence to single-inhaler triple therapy (SITT) versus MITT in a real-world setting in England.Patients and MethodsThis was a retrospective cohort study using linked primary care (Clinical Practice Research Datalink Aurum) and secondary care (Hospital Episode Statistics [HES] Admitted Patient Care) data to identify patients with COPD who were newly initiated on SITT or MITT between November 2017 and June 2019. Eligible patients were aged ≥35 years and had a forced expiratory volume in 1 second/forced vital capacity <0.7, linkage to HES and continuous registration with a general practitioner for 12 months pre- and 6 months post-initiation. Inverse probability of treatment weighting was used to balance baseline characteristics between cohorts. Adherence was measured using the proportion of days covered by days’ supply of SITT or MITT prescriptions. Persistence was measured with a gap of >30 days between the end of a prescription and the following refill used to determine non-persistence.ResultsOverall, 4080 SITT and 6579 MITT users comprised the study cohort. After weighting, the baseline characteristics between the cohorts were comparable (absolute standardized mean difference <10%). SITT users had significantly higher adherence than MITT users at 6, 12, and 18 months post-initiation (p<0.001 for all comparisons). Median persistence was higher among SITT users than MITT users (5.09 months vs 0.99 months).ConclusionPatients with COPD in England initiating SITT had significantly better adherence and persistence compared with MITT initiators. These improvements continued at least 18 months following treatment initiation.

Very high PD-L1 expression as a prognostic indicator of overall survival among patients with advanced non-small cell lung cancer receiving anti-PD-(L)1 monotherapies in routine practice.

Programmed death or ligand-1 (PD-(L)1) pathway inhibitors confer improved survival as the first-line treatment for advanced non-small cell lung cancer (aNSCLC) in patients with PD-L1 expression (PD-L1 + e ≥ 50%) compared to platinum-doublet chemotherapy and have become a standard therapy. Some recent evidence suggests that among aNSCLC patients with PD-L1 + e of ≥50% receiving pembrolizumab monotherapy, very high levels of PD-L1 + e (≥90%) may be associated with better outcomes. We sought to assess whether very high PD-L1 + e (≥90%) compared to high PD-L1 + e (50%-89%) is associated with an overall survival benefit in aNSCLC patients receiving anti-PD-(L)1 monotherapies. We conducted a single-site retrospective cohort study of aNSCLC patients who initiated PD-(L)1 inhibitor monotherapy as the first-line treatment from October 24, 2016, to August 25, 2021, and had a PD-L1 + e ≥ 50%. The primary outcome was overall survival, measured from the start of the first-line PD-(L)1 inhibitor monotherapy (index date) to date of death or last confirmed activity prior to the cohort exit date. Propensity score-based inverse probability weighting (IPW) was used to control for confounding in Kaplan-Meier curves and Cox proportional hazard regression analysis. One hundred sixty-six patients with aNSCLC receiving PD-(L)1 inhibitor monotherapy met inclusion criteria. 54% were female, 90% received pembrolizumab, median age was 68 years, 70% had non-squamous cell carcinoma, 94% had a history of smoking, 29% had a KRAS mutation, and 37% had very high PD-L1 + e. Unweighted covariates at cohort entry were similar between groups (absolute standardized mean differences [SMDs] <0.1) except for race (SMD=0.2); age at therapy initiation (SMD=0.13); smoking status (SMD=0.13), and BRAF mutation status (SMD=0.11). After weighting, baseline covariates were well balanced (all absolute SMDs <0.1). In the weighted analysis, having a very high PD-L1 + e was associated with lower mortality (weighted hazard ratio 0.57, 95% CI 0.36-0.90) and longer median survival: 3.85 versus 1.49 years. Very high PD-L1 + e (≥90%) was associated with an overall survival benefit over high PD-L1 + e (50%-89%) in patients receiving the first-line PD-(L)1 inhibitor monotherapy in a model controlling for potential confounders. These findings should be confirmed in a larger real-world data set.

Identifying and handling unbalanced baseline characteristics in a non-randomized, controlled, multicenter social care nurse intervention study for patients in advanced stages of cancer

PurposeGiven the psychosocial burdens patients in advanced stages of cancer face, innovative care concepts are needed. At the same time, such vulnerable patient groups are difficult to reach for participation in intervention studies and randomized patient inclusion may not be feasible. This article aims to identify systematic biases respectively selection effects occurring during the recruitment phase and to discuss their potential causes based on a non-randomized, multicenter intervention study with patients in advanced stages of cancer.MethodsPatients diagnosed with at least one of 16 predefined cancers were recruited at four hospitals in three German cities. The effect of social care nurses’ continuous involvement in acute oncology wards was measured by health-related quality of life (EORTC QLQ-C30), information and participation preferences, decisional conflicts, doctor-patient communication, health literacy and symptom perception. Absolute standardized mean difference was calculated as a standardized effect size to test baseline characteristics balance between the intervention and control groups.ResultsThe study enrolled 362 patients, 150 in the intervention and 212 in the control group. Except for gender, both groups differed in relevant socio-demographic characteristics, e.g. regarding age and educational background. With respect to the distribution of diagnoses, the intervention group showed a higher symptom burden than the control group. Moreover, the control group reported better quality of life at baseline compared to the intervention group (52.6 points (SD 21.7); 47.8 points (SD 22.0), ASMD = 0.218, p = 0.044).ConclusionOverall, the intervention group showed more social and health vulnerability than the control group. Among other factors, the wide range of diagnoses included and structural variation between the recruiting clinics increased the risk for bias. We recommend a close, continuous monitoring of relevant social and health-related characteristics during the recruitment phase as well as the use of appropriate statistical analysis strategies for adjustment, such as propensity score methods.Trial registration: German Clinical Trials Register (DRKS-ID: DRKS00013640); registered on 29th December 2017.

The value of combining individual and small area sociodemographic data for assessing and handling selective participation in cohort studies: Evidence from the Swedish CardioPulmonary bioImage Study.

To study the value of combining individual- and neighborhood-level sociodemographic data to predict study participation and assess the effects of baseline selection on the distribution of metabolic risk factors and lifestyle factors in the Swedish CardioPulmonary bioImage Study (SCAPIS). We linked sociodemographic register data to SCAPIS participants (n = 30,154, ages: 50-64 years) and a random sample of the study's target population (n = 59,909). We assessed the classification ability of participation models based on individual-level data, neighborhood-level data, and combinations of both. Standardized mean differences (SMD) were used to examine how reweighting the sample to match the population affected the averages of 32 cardiopulmonary risk factors at baseline. Absolute SMDs >0.10 were considered meaningful. Combining both individual-level and neighborhood-level data gave rise to a model with better classification ability (AUC: 71.3%) than models with only individual-level (AUC: 66.9%) or neighborhood-level data (AUC: 65.5%). We observed a greater change in the distribution of risk factors when we reweighted the participants using both individual and area data. The only meaningful change was related to the (self-reported) frequency of alcohol consumption, which appears to be higher in the SCAPIS sample than in the population. The remaining risk factors did not change meaningfully. Both individual- and neighborhood-level characteristics are informative in assessing study selection effects. Future analyses of cardiopulmonary outcomes in the SCAPIS cohort can benefit from our study, though the average impact of selection on risk factor distributions at baseline appears small.

Abstract P2-11-05: Generating real-world external comparators for randomized clinical trials (RCTs) in metastatic breast cancer (mBC) using electronic health records (EHRs)

Abstract Objectives: Real-world data (RWD) from the routine care of patients with cancer captured through EHRs is a valuable resource for research. Understanding the relationship between characteristics and outcomes of patients treated in the real world and those treated in clinical trials is essential to produce evaluable trial-like populations using RWD in oncology for research and regulatory purposes. Methods: This study used: a) RWD from the Flatiron Health EHR-derived, de-identified, longitudinal database (comprising patient-level structured and unstructured data, curated via technology-enabled abstraction selected from approximately 280 US cancer clinics [~800 sites of care]) and b) patient-level data from three completed RCTs (PALOMA-2, MONALEESA-2, and MONARCH-3) including patients with previously untreated hormone receptor positive (HR+), HER2/neu negative (HER2-) mBC, then separately pooled across the trials into two treatment groups, patients who received aromatase inhibitor monotherapy (AI) or a CDK4/6 inhibitor + AI. Key eligibility criteria were similar across the RCTs and were used to select a real world external cohort (rwEC) initiating AI monotherapy on or prior to 11 Nov 2015 (end of MONARCH-3 enrollment period). Patients from the rwEC were matched separately to the control arm and experimental arm patients from the pooled RCT using propensity score method (PSM). The propensity score was estimated by a logistic regression using baseline covariates of age, race, site of disease (visceral, non-visceral), Eastern Cooperative Oncology Group Performance Status (ECOG PS) (0, 1), and metastatic disease (recurrent, new). The matching ratio was 1:1 without replacement with calipers. Covariate balance was measured by the absolute standardized mean difference (ASMD). Due to the high percentage of missing ECOG PS data, matching was repeated 100 times with imputed ECOG PS. The impact of including additional key covariates for propensity matching such as number of disease sites, bone-only disease, and prior endocrine therapy was assessed. Results: There were 1326 patients with HR+, HER2- mBC selected from the EHR-derived database who received first-line AI therapy and 1827 patients randomized in the RCTs (1106 and 721 patients for experimental and control arms, respectively). With 100 matching iterations, 563 rwEC patients on average (range, 547-572) were matched to the RCTs control arm, and 753 rwEC patients on average (range: 741-761) were matched to the RCTs experimental arm. Prior to matching, the ASMD varied widely across all prespecified baseline covariates (4.3 for the rwEC vs. RCTs control arm, 2.6 for the rwEC vs. RCTs experimental arm). After matching was performed, across all baseline covariates used in the PSM, the ASMD was reduced to be under 0.12 for the rwEC vs. RCTs control arm, and under 0.2 for the rwEC vs. RCTs experimental arm in more than 90% of the matching iterations. Analyses looking at the additional baseline covariates to the propensity matching resulted in similar ASMDs. Conclusions: EHR-derived RWD can be used to generate a cohort of patients with similar baseline characteristics to those treated on RCT. The next step in our trial emulation framework is to analyze the comparability of outcomes between these two matched cohorts. Citation Format: Laleh Amiri-Kordestani, Xin Gao, Shrujal Baxi, Erik Bloomquist, Jonathan Bryan, Lynn Howie, Catherine Keane, Paul G. Kluetz, Christy Osgood, Prashni Paliwal, Donna R. Rivera, James Roose, Julie Schneider, Harpreet Singh, Shenghui Tang, Lijun Zhang, Julia A. Beaver. Generating real-world external comparators for randomized clinical trials (RCTs) in metastatic breast cancer (mBC) using electronic health records (EHRs) [abstract]. In: Proceedings of the 2021 San Antonio Breast Cancer Symposium; 2021 Dec 7-10; San Antonio, TX. Philadelphia (PA): AACR; Cancer Res 2022;82(4 Suppl):Abstract nr P2-11-05.