Greedy caliper propensity score matching can yield variable estimates of the treatment-outcome association-A simulation study.

PURPOSE: Greedy caliper propensity score (PS) matching is dependent on randomness, which can ultimately affect causal estimates. We sought to investigate the variation introduced by this randomness. METHODS: Based on a literature search to define the simulation parameters, we simulated 36 cohorts of different sizes, treatment prevalence, outcome prevalence, treatment-outcome-association. We performed 1:1 caliper and nearest neighbor (NN) caliper PS-matching and repeated this 1000 times in the same cohort, before calculating the treatment-outcome association. RESULTS: Repeating caliper and NN caliper matching in the same cohort yielded large variations in effect estimates, in all 36 scenarios, with both types of matching. The largest variation was found in smaller cohorts, where the odds ratio (OR) ranged from 0.53 to 10.00 (IQR of ORs: 1.11-1.67). The 95% confidence interval was not consistently overlapping a neutral association after repeating the matching with both algorithms. We confirmed these findings in a noninterventional example study. CONCLUSION: Caliper PS-matching can yield highly variable estimates of the treatment-outcome association if the analysis is repeated.

Determining the effect of one decade on fitness of elite Austrian youth soccer players using propensity score matching.

Current trends in attacking strategies and increases in external workload have led to a need for fast and well-conditioned athletes in modern soccer. More recently, progressions in speed, coordination, power and endurance were found over a decade in elite Austrian youth players. However, possible confounders such as relative age, maturation, learning effects, and academy philosophy may have influenced these changes. The present study aimed to determine the decade effect on fitness under statistical control of players' exact age, height, body mass, test location as well as total number of pretests and time interval between test and pretest. Players annually completed a battery of anthropometric, general and soccer-specific fitness tests. MANCOVA was calculated to identify the overall impacts of the covariates on fitness. To balance the covariates of initially 2,530 "former" (2002 to 2005) and 2,611 "recent" (2012 to 2015) players, 1:1 nearest neighbor propensity score (PS) matching was used, resulting in 587 U13, 573 U14, 475 U15, 325 U16, 262 U17, and 129 U18 matched pairs. The decade effect on fitness was assessed by independent t-tests and Cohen's d separately at each age group. Superior performances of recent players were found for linear sprint across all age categories (d = 0.154-0.476) as well as for agility (d = 0.125-0.340) and change-of-direction speed (d = 0.172-0.466) in U15 to U18. Reaction speed increased in U13 (d = 0.288) and U15 (d = 0.310). Flexibility reduced over the decade in all age categories (d = -0.151 to -0.589) and upper-limb power decreased (d = -0.278 to -0.347) in U13 and U14. Balancing the covariate distribution via PS matching generally confirmed previous findings, with fitness decade effects reflecting the athletic needs for modern soccer. Since fitness performance changed over time, reference values should be periodically updated. Coaches favor both physical and cognitive fast players nowadays. Thus, training should target all aspects of speed, without disregarding flexibility, upper-limb power and other preventive strategies that keep the players on the pitch.

Evaluation of a Home-Based, Nurse Practitioner-led Advanced Illness Care Program.

OBJECTIVES: In the United States, people with serious illness often experience gaps and discontinuity in care. Gaps are frequently exacerbated by limited mobility, need for social support, and challenges managing multiple comorbidities. The Advanced Illness Care (AIC) Program provides nurse practitioner-led, home-based care for people with serious or complex chronic illnesses that specifically targets palliative care needs and coordinates with patients' primary care and specialty health care providers. We sought to investigate the effect of the AIC Program on hospital encounters [hospitalizations and emergency department (ED) visits], hospice conversion, and mortality. DESIGN: Retrospective nearest-neighbor matching. SETTING AND PARTICIPANTS: Patients in AIC who had ≥1 inpatient stay within the 60 days prior to AIC enrollment to fee-for-service Medicare controls at 9 hospitals within one health system. METHODS: We matched on demographic characteristics and comorbidities, with exact matches for diagnosis-related group and home health enrollment. Outcomes were hospital encounters (30- and 90-day ED visits and hospitalizations), hospice conversion, and 30- and 90-day mortality. RESULTS: We included 110 patients enrolled in the AIC Program matched to 371 controls. AIC enrollees were mean age 77.0, 40.9% male, and 79.1% white. Compared with controls, AIC enrollees had a higher likelihood of ED visits at 30 [15.1 percentage points, confidence interval (CI) 4.9, 25.3; P = .004] and 90 days (27.8 percentage points, CI 16.0, 39.6; P < .001); decreased likelihood of hospitalization at 30 days (11.4 percentage points, CI -17.7, -5.0; P < .001); and a higher likelihood of converting to hospice (22.4 percentage points, CI 11.4, 33.3; P < .001). CONCLUSIONS: The AIC Program provides care and coordination that the home-based serious illness population may not otherwise receive. IMPLICATIONS: By identifying and addressing care needs and gaps in care early, patients may avoid unnecessary hospitalizations and receive timely hospice services as they approach the end of life.

High-risk myeloma and minimal residual disease postautologous-HSCT predict worse outcomes.

The aim of our study was to determine the impact of high-risk disease (HRD) and MRD on outcomes in myeloma patients receiving bortezomib-based induction followed by autologous hematopoietic stem cell transplant (auto-HSCT). HRD included t(4:14), t(14;16), del 17p, del 1p and/or amplification 1q by cytogenetics/FISH; all others were standard-risk disease (SRD). A subset of 165 newly diagnosed myeloma patients in a 2:1 ratio of HRD:SRD was generated using propensity score based nearest neighbor matching. Multiparametric flow cytometry (MFC) was used to detect MRD after auto-HSCT in select patients. MRD+ status at 3 months post auto-HSCT (hazard ratio (HR = 4.23, p = .028) and HRD (HR = 1.72, p = .026) were associated with a shorter PFS. Similarly, MRD+ 3 months post auto-HSCT (HR = 6.93, p = .08) and HRD (HR = 3.54, p < .001) and were associated with a shorter OS. Despite bortezomib-based induction, upfront auto-HSCT, and use of maintenance therapy, PFS and OS remained worse in MRD+ and HRD patients.

Identifying compliant participants through data matching improved estimation of intervention efficacy: randomized trials with opt-in/opt-out strategies.

OBJECTIVES: We propose a data-matching approach to estimate intervention efficacy for randomized controlled trials (RCTs) when there is noncompliance to the allocated treatment with induced selection bias. STUDY DESIGN AND SETTING: We considered a large RCT to compare health care costs and hospital length of stay 12 months after randomization. Participants allocated to the intervention group were eligible to receive health-coaching and disease-management services. An opt-out approach was adopted for recruitment. Control-group participants received usual care but were allowed to opt-in to receive the intervention. Using "nearest-neighbor"-matched data, we identified compliant participants in both arms to estimate intervention efficacy. Results were compared with intention-to-treat (ITT), instrumental-variable-adjusted ITT, per-protocol (PP), and as-treated (AT) analyses. RESULTS: The ITT estimated an intervention effect of a 1.5% reduction in cost, but 56.7% of intervention-group participants did not receive health coaching. The PP and AT found an increase in cost of 9.4% and 17.1%, respectively. The matching method estimated a 12.3% reduction in cost. After adjustment for baseline covariates, the intervention group had lower same-day admission cost (13.6%; 95% CI: 7.3%-20.0%; P < 0.001) and shorter hospital stay (11.2%; 95% CI: 2.6%-19.9%; P = 0.021). CONCLUSION: Opt-in/opt-out strategies in RCTs misled intervention comparisons and the matching approach improved estimation of intervention efficacy.

Effects of neoadjuvant chemotherapy combined with radiotherapy in patients with advanced esophageal carcinoma.

We analyzed the improvement of survival time and the effects of neoadjuvant chemotherapy combined with radiotherapy on treating patients with advanced esophageal carcinoma. Retrospectively, 43 patients were selected with esophageal carcinoma who were administered neoadjuvant chemotherapy combined with radiotherapy. According to gender, and tumor staging, the nearest neighbor matching was carried out. Eighty-six patients (1:2) who received neoadjuvant chemotherapy and 129 patients (1:3) who underwent surgery only were taken and compared for clinical outcomes. It was found that in the combination group, the median survival time was prolonged and the 1-year survival rate improved. The diameter of tumors was significantly reduced, and the surgical resection, margin negative and total effective rates improved. In addition, the recurrence rate significantly decreased, whereas quality of life scores significantly increased (p<0.05). The comparison of overall incidence of complications was not statistically significant (p>0.05). Tumor staging, location, and diameter after neoadjuvant therapy, as well as therapeutic regimen, treatment cycle, margin negative rate and effective rate were independent risk factors for significantly influencing survival outcomes and time (p<0.05). In conclusion, neoadjuvant chemotherapy combined with radiotherapy can be utilized to treat advanced esophageal carcinoma improve survival time and promote prognosis.