Category: Statistics

Am J Manag Care. 2023 Apr 1;29(4):e124-e128. doi: 10.37765/ajmc.2023.89353.

ABSTRACT

OBJECTIVES: We aimed to describe the experience of a state Medicaid agency incentivizing reduction of racial and ethnic disparities in a hospital quality incentive program (QIP).

STUDY DESIGN: Retrospective review of a decade of experience implementing a hospital health disparity (HD) composite measure.

METHODS: Observational analysis of programwide trends in missed opportunity rates and between-group variance (BGV) for the HD composite from 2011 to 2020 and subanalysis of 16 metrics included in the HD composite for at least 4 years over the decade.

RESULTS: Programwide missed opportunity rates and BGV fluctuated widely from 2011 to 2020, likely due to variation in measures included in the HD composite. When the 16 measures that were included in the HD composite for at least 4 years were collapsed into a hypothetical 4-year period, missed opportunity rates decreased across the 4 consecutive years, from 47% in year 1 to 20% in year 4. Differences among racial and ethnic subgroups also decreased across the 4-year period, as reflected in the BGV decrease from 7.85 × 10-4 in year 1 to 5.10 × 10-4 in year 4.

CONCLUSIONS: Construction of a composite measure, use of a summary disparity statistic, and measure selection are key considerations in the design and interpretation of equity-focused payment programs. This analysis revealed improved aggregate quality performance and a modest reduction in racial and ethnic disparities for measures included in the HD composite for at least 4 years. Further research is needed to evaluate the association between equity-oriented incentives and health disparities.

PMID:37104839 | DOI:10.37765/ajmc.2023.89353

Am J Manag Care. 2023 Apr 1;29(4):e117-e123. doi: 10.37765/ajmc.2023.89352.

ABSTRACT

OBJECTIVES: To determine whether broad categories of criteria exist among prior authorization (PA) policies from different managed care organizations (MCOs) and to identify similarities and differences among MCO coverage requirements for medications within the calcitonin gene-related peptide (CGRP) antagonist class.

STUDY DESIGN: Quantitative and qualitative descriptive analysis.

METHODS: PA policies from different MCOs for erenumab, fremanezumab, galcanezumab, and eptinezumab were identified through a comprehensive online search. Individual criteria from each policy were analyzed and grouped into both broad and specific categories. Descriptive statistics were used to identify and summarize trends among policies.

RESULTS: A total of 47 MCOs were included in the analysis. The vast majority of policies applied to galcanezumab (n = 45; 96%), erenumab (n = 44; 94%), and fremanezumab (n = 40; 85%), with fewer policies for eptinezumab (n = 11; 23%). There were 5 broad categories of PA criteria found to be included in coverage policies: prescriber specialization (n = 21; 45%), prerequisite drugs (n = 45; 96%), safety considerations (n = 8; 17%), and response to therapy (n = 43; 91%). The final category, titled appropriate use, included any criteria meant to ensure appropriate medication use and included age requirements (n = 26; 55%), suitable diagnosis (n = 34; 72%), exclusion of other diagnoses (n = 17; 36%), and exclusion of concurrent medications (n = 22; 47%).

CONCLUSIONS: This study identified 5 broad categories of PA criteria used by MCOs in the management of CGRP antagonists. However, within these categories, specific criteria from different MCOs varied significantly.

PMID:37104838 | DOI:10.37765/ajmc.2023.89352

J Exp Psychol Gen. 2023 Apr 27. doi: 10.1037/xge0001357. Online ahead of print.

ABSTRACT

Researchers often remove outliers when comparing groups. It is well documented that the common practice of removing outliers within groups leads to inflated Type I error rates. However, it was recently argued by André (2022) that if outliers are instead removed across groups, Type I error rates are not inflated. The same study discusses that removing outliers across groups is a specific case of the more general concept of hypothesis-blind removal of outliers, which is consequently recommended. In this paper, I demonstrate that, contrary to this advice, hypothesis-blind outlier removal is problematic. Specifically, it almost always invalidates confidence intervals and biases estimates if there are group differences. It moreover inflates Type I error rates in certain situations, for example, when variances are unequal and data nonnormal. Consequently, a data point may not be removed solely because it is deemed an outlier, whether the procedure used is hypothesis-blind or hypothesis-aware. I conclude by recommending valid alternatives. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

PMID:37104797 | DOI:10.1037/xge0001357

Clin Orthop Relat Res. 2023 Apr 27. doi: 10.1097/CORR.0000000000002627. Online ahead of print.

ABSTRACT

BACKGROUND: The initial approach to the treatment of desmoid tumors has changed from surgical resection to watchful waiting. However, surgery is still sometimes considered for some patients, and it is likely that a few patients would benefit from tumor removal if the likelihood of local recurrence could be predicted. However, to our knowledge, there is no tool that can provide guidance on this for clinicians at the point of care.

QUESTION/PURPOSE: We sought to explore whether a combined molecular and clinical prognostic model for relapse in patients with desmoid tumors treated with surgery would allow us to identify patients who might do well with surgical excision.

METHODS: This was a retrospective, single-center study of 107 patients with desmoid tumors who were surgically treated between January 1980 and December 2015, with a median follow-up of 106 months (range 7 to 337 months). We correlated clinical variables (age, tumor size, and localization) and CTNNB1 gene mutations with recurrence-free survival. Recurrence-free survival was estimated using a Kaplan-Meier curve. Univariate and multivariable analyses of time to local recurrence were performed using Cox regression models. A final nomogram model was constructed according to the final fitted Cox model. The predictive performance of the model was evaluated using measures of calibration and discrimination: calibration plot and the Harrell C-statistic, also known as the concordance index, in which values near 0.5 represent a random prediction and values near 1 represent the best model predictions.

RESULTS: The multivariable analysis showed that S45F mutations (hazard ratio 5.25 [95% confidence interval 2.27 to 12.15]; p < 0.001) and tumor in the extremities (HR 3.15 [95% CI 1.35 to 7.33]; p = 0.008) were associated with a higher risk of local recurrence. Based on these risk factors, we created a model; we observed that patients considered to be at high risk of local recurrence as defined by having one or two factors associated with recurrence (extremity tumors and S45F mutation) had an HR of 8.4 compared with patients who had no such factors (95% CI 2.84 to 24.6; p < 0.001). From these data and based on the multivariable Cox models, we also developed a nomogram to estimate the individual risk of relapse after surgical resection. The model had a concordance index of 0.75, or moderate discrimination.

CONCLUSION: CTNNB1 S45F mutations combined with other clinical variables are a potential prognostic biomarker associated with the risk of relapse in patients with desmoid tumors. The developed nomogram is simple to use and, if validated, could be incorporated into clinical practice to identify patients at high risk of relapse among patients opting for surgical excision and thus help clinicians and patients in decision-making. A large multicenter study is necessary to validate our model and explore its applicability.

LEVEL OF EVIDENCE: Level III, therapeutic study.

PMID:37104792 | DOI:10.1097/CORR.0000000000002627

J Pediatr Orthop. 2023 Apr 27. doi: 10.1097/BPO.0000000000002420. Online ahead of print.

ABSTRACT

BACKGROUND: Patella alta is an anatomic risk factor for patellar instability in adolescents that is also linked to the risk factor of trochlear dysplasia. This study aims to determine the age of onset and age-related incidence of patella alta in a pediatric population of patients with patellar instability. We hypothesized that patellar height ratios would not increase with age, suggesting a congenital rather than the developmental origin of patella alta.

METHODS: A retrospective cross-sectional cohort of patients was collected with the following inclusion criteria: patients aged 5 to 18 who had a knee magnetic resonance imaging performed from 2000 to 2022 and the International Classification of Diseases code for patellar dislocation. Demographic information and details of the patellar instability episode(s) were collected with a chart review. Sagittal magnetic resonance imaging was used to measure Caton-Deschamps Index (CDI) and the Insall-Salvati Ratio (ISR) by 2 observers. Data were analyzed to assess for associations between patellar height ratios and age of the first dislocation and to assess if the proportion of patients categorized as having patella alta changed with age.

RESULTS: The 140 knees included in the cohort had an average age of 13.9 years (SD=2.40; range: 8-18) and were 55% female. Patella alta was present in 78 knees (55.7%) using CDI>=1.2 and in 59 knees (42.1%) using ISR>=1.3. The earliest age patella alta was observed was at age 8 using CDI>=1.2 and age 10 using ISR>=1.3. There were no statistically significant associations between CDI and age without adjustment (P=0.14) nor after adjustment for sex and body mass index (P=0.17). The proportion of knees above the CDI threshold for patella alta to the knees below the cutoff did not show a significant change with age (P=0.09).

CONCLUSIONS: Patella alta, as defined by CDI, is seen in patients as young as 8 years old. Patellar height ratios do not change with age in patients with patellar dislocation, suggesting that patella alta is established at a young age rather than developing during the adolescent years.

LEVEL OF EVIDENCE: Level III-diagnostic, cross-sectional.

PMID:37104788 | DOI:10.1097/BPO.0000000000002420

Psychol Methods. 2023 Apr 27. doi: 10.1037/met0000573. Online ahead of print.

ABSTRACT

Testing null hypotheses of the form “β = 0,” by the use of various Null Hypothesis Significance Tests (rendering a dichotomous reject/not reject decision), is considered standard practice when evaluating the individual parameters of statistical models. Bayes factors for testing these (and other) hypotheses allow users to quantify the evidence in the data that is in favor of a hypothesis. Unfortunately, when testing equality-contained hypotheses, the Bayes factors are sensitive to the specification of prior distributions, which may be hard to specify by applied researchers. The paper proposes a default Bayes factor with clear operating characteristics when used for testing whether the fixed parameters of linear two-level models are equal to zero. This is achieved by generalizing an already existing approach for linear regression. The generalization requires: (a) the sample size for which a new estimator for the effective sample size in two-level models containing random slopes is proposed; (b) the effect size for the fixed effects for which the so-called marginal R² for the fixed effects is used. Implementing the aforementioned requirements in a small simulation study shows that the Bayes factor yields clear operating characteristics regardless of the value for sample size and the estimation method. The paper gives practical examples and access to an easy-to-use wrapper function to calculate Bayes factors for hypotheses with respect to the fixed coefficients of linear two-level models by using the R package bain. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

PMID:37104765 | DOI:10.1037/met0000573

Psychol Methods. 2023 Apr 27. doi: 10.1037/met0000556. Online ahead of print.

ABSTRACT

Likelihood ratio tests (LRTs) are a popular tool for comparing statistical models. However, missing data are also common in empirical research, and multiple imputation (MI) is often used to deal with them. In multiply imputed data, there are multiple options for conducting LRTs, and new methods are still being proposed. In this article, we compare all available methods in multiple simulations covering applications in linear regression, generalized linear models, and structural equation modeling. In addition, we implemented these methods in an R package, and we illustrate its application in an example analysis concerned with the investigation of measurement invariance. (PsycInfo Database Record (c) 2023 APA, all rights reserved).

PMID:37104764 | DOI:10.1037/met0000556

J Clin Oncol. 2023 Apr 27:JCO2202390. doi: 10.1200/JCO.22.02390. Online ahead of print.

ABSTRACT

PURPOSE: It remains unknown whether or not short-term androgen deprivation (STAD) improves survival among men with intermediate-risk prostate cancer (IRPC) treated with dose-escalated radiotherapy (RT).

METHODS: The NRG Oncology/Radiation Therapy Oncology Group 0815 study randomly assigned 1,492 patients with stage T2b-T2c, Gleason score 7, or prostate-specific antigen (PSA) value >10 and ≤20 ng/mL to dose-escalated RT alone (arm 1) or with STAD (arm 2). STAD was 6 months of luteinizing hormone-releasing hormone agonist/antagonist therapy plus antiandrogen. RT modalities were external-beam RT alone to 79.2 Gy or external beam (45 Gy) with brachytherapy boost. The primary end point was overall survival (OS). Secondary end points included prostate cancer-specific mortality (PCSM), non-PCSM, distant metastases (DMs), PSA failure, and rates of salvage therapy.

RESULTS: Median follow-up was 6.3 years. Two hundred nineteen deaths occurred, 119 in arm 1 and 100 in arm 2. Five-year OS estimates were 90% versus 91%, respectively (hazard ratio [HR], 0.85; 95% CI, 0.65 to 1.11]; P = .22). STAD resulted in reduced PSA failure (HR, 0.52; P <.001), DM (HR, 0.25; P <.001), PCSM (HR, 0.10; P = .007), and salvage therapy use (HR, 0.62; P = .025). Other-cause deaths were not significantly different (P = .56). Acute grade ≥3 adverse events (AEs) occurred in 2% of patients in arm 1 and in 12% for arm 2 (P <.001). Cumulative incidence of late grade ≥3 AEs was 14% in arm 1 and 15% in arm 2 (P = .29).

CONCLUSION: STAD did not improve OS rates for men with IRPC treated with dose-escalated RT. Improvements in metastases rates, prostate cancer deaths, and PSA failures should be weighed against the risk of adverse events and the impact of STAD on quality of life.

PMID:37104748 | DOI:10.1200/JCO.22.02390

Pain Med. 2023 Apr 27:pnad049. doi: 10.1093/pm/pnad049. Online ahead of print.

ABSTRACT

OBJECTIVE: To investigate how a behavioral health, artificial intelligence (AI)-powered, digital self-management tool affects the daily functions in adults with chronic back and neck pain.

DESIGN: Eligible subjects were enrolled in a 12-week prospective, multicenter, single-arm, open-label study and instructed to use the digital coach daily. Primary outcome was a change in Patient-Reported Outcomes Measurement Information Systems (PROMIS) scores for pain interference. Secondary outcomes were changes in PROMIS physical function, anxiety, depression, pain intensity scores and pain catastrophizing scale (PCS) scores.

METHODS: Subjects logged daily activities, using PainDrainerTM, and data analyzed by the AI engine. Questionnaire and web-based data were collected at 6 and 12-weeks and compared to subjects’ baseline.

RESULTS: Subjects completed the 6- (n = 41) and 12-week (n = 34) questionnaires. A statistically significant Minimal Important Difference (MID) for pain interference was demonstrated in 57.5% of the subjects. Similarly, MID for physical function was demonstrated in 72.5% of the subjects. A pre- to post-intervention improvement in depression score was also statistically significant, observed in 100% of subjects, as was the improvement in anxiety scores, evident in 81.3% of the subjects. PCS mean scores was also significantly decreased at 12 weeks.

CONCLUSION: Chronic pain self-management, using an AI-powered, digital coach anchored in behavioral health principles significantly improved subjects’ pain interference, physical function, depression, anxiety, and pain catastrophizing over the 12-week study period.

PMID:37104747 | DOI:10.1093/pm/pnad049

Bioinformatics. 2023 Apr 27:btad291. doi: 10.1093/bioinformatics/btad291. Online ahead of print.

ABSTRACT

MOTIVATION: Testing association between multiple phenotypes with a set of genetic variants simultaneously, rather than analyzing one trait at a time, is receiving an increasing attention for its high statistical power and easy explanation on pleiotropic effects. The kernel-based association test (KAT), being free of data dimensions and structures, has proven to be a good alternative method for genetic association analysis with multiple phenotypes. However, KAT suffers from substantial power loss when multiple phenotypes have moderate to strong correlations. To handle this issue, we propose a maximum kernel-based association test (MaxKAT) and suggest using the generalized extreme value distribution to calculate its statistical significance under the null hypothesis.

RESULTS: We show that MaxKAT reduces computational intensity greatly while maintaining high accuracy. Extensive simulations demonstrate that MaxKAT can properly control type I error rates and obtain remarkably higher power than KAT under most of the considered scenarios. Application to a porcine dataset used in biomedical experiments of human disease further illustrates its practical utility.

AVAILABILITY AND IMPLEMENTATION: The R package MaxKAT that implements the proposed method is available on Github https://github.com/WangJJ-xrk/MaxKAT.

PMID:37104737 | DOI:10.1093/bioinformatics/btad291