Concave 1-norm group selection.

Grouping structures arise naturally in many high-dimensional problems. Incorporation of such information can improve model fitting and variable selection. Existing group selection methods, such as the group Lasso, require correct membership. However, in practice it can be difficult to correctly specify group membership of all variables. Thus, it is important to develop group selection methods that are robust against group mis-specification. Also, it is desirable to select groups as well as individual variables in many applications. We propose a class of concave [Formula: see text]-norm group penalties that is robust to grouping structure and can perform bi-level selection. A coordinate descent algorithm is developed to calculate solutions of the proposed group selection method. Theoretical convergence of the algorithm is proved under certain regularity conditions. Comparison with other methods suggests the proposed method is the most robust approach under membership mis-specification. Simulation studies and real data application indicate that the [Formula: see text]-norm concave group selection approach achieves better control of false discovery rates. An R package grppenalty implementing the proposed method is available at CRAN.

Geographic patterns in patient demographics and insulin use in 18 countries, a global perspective from the multinational observational study assessing insulin use: understanding the challenges associated with progression of therapy (MOSAIc).

BACKGROUND: Among patients with type 2 diabetes, insulin intensification to achieve glycemic targets occurs less often than clinically indicated. Barriers to intensification are not well understood. We present patients' baseline characteristics from MOSAIc, a study investigating patient-, physician-, and healthcare environment-based factors affecting insulin intensification and subsequent health outcomes. METHODS: MOSAIc is a longitudinal, observational study following patients' diabetes care in 18 countries: United Arab Emirates (UAE), Argentina, Brazil, Canada, China, Germany, India, Israel, Italy, Japan, Mexico, Russia, Saudi Arabia, South Korea, Spain, Turkey, United Kingdom, United States. Eligible patients are age ≥ 18, have type 2 diabetes, and have used insulin for ≥ 3 months with/without other antidiabetic medications. Extensive baseline demographic, clinical, and psychosocial data are collected at baseline and regular intervals during the 24-month follow-up. We conducted descriptive analyses of baseline data. RESULTS: Four thousand three hundred forty one patients met eligibility criteria. Patients received their type 2 diabetes diagnosis 12 ± 8 years prior to baseline visit, yet patients in developing countries were younger than in developed countries (e.g., UAE, 55 ± 10; Germany = 70 ± 10). Saudi Arabians had the highest HbA1c values (9.0 ± 2.2) and Germany (7.5 ± 1.4) among the lowest. Most patients in 5 (28%) of the 18 countries did not use an oral antidiabetic drug. Over half of patients in fourteen (78 %) countries exclusively used basal insulin; most Indian and Chinese patients exclusively used mixed insulin. CONCLUSIONS: MOSAIc's baseline data highlight differences in patient characteristics across countries. These patterns, along with physician and healthcare environment differences, may contribute to the likelihood of insulin intensification and subsequent clinical outcomes.

Treatment Discontinuation Patterns for Patients With Chronic Lymphocytic Leukemia in Real-World Settings: Results From a Multi-Center International Study.

INTRODUCTION: This study assessed treatment discontinuation patterns and reasons among chronic lymphocytic leukemia (CLL) patients initiating first-line (1L) and second-line (2L) treatments in real-world settings. MATERIALS AND METHODS: Using deidentified electronic medical records from the CLL Collaborative Study of Real-World Evidence, premature treatment discontinuation was assessed among FCR, BR, BTKi-based, and BCL-2-based regimen cohorts. RESULTS: Of 1364 1L patients (initiated in 1997-2021), 190/13.9% received FCR (23.7% discontinued prematurely); 255/18.7% received BR (34.5% discontinued prematurely); 473/34.7% received BTKi-based regimens, of whom 28.1% discontinued prematurely; and 43/3.2% received venetoclax-based regimens, of whom 16.3% discontinued prematurely (venetoclax monotherapy: 7/0.5%, of whom 42.9% discontinued; VG/VR: 36/2.6%, of whom 11.1% discontinued). The most common reasons for treatment discontinuation were adverse events (FCR: 25/13.2%; BR: 36/14.1%; BTKi-based regimens: 75/15.9%) and disease progression (venetoclax-based: 3/7.0%). Of 626 2L patients, 20/3.2% received FCR (50.0% discontinued); 62/9.9% received BR (35.5% discontinued); 303/48.4% received BTKi-based regimens, of whom 38.0% discontinued; and 73/11.7% received venetoclax-based regimens, of whom 30.1% discontinued (venetoclax monotherapy: 27/4.3%, of whom 29.6% discontinued; VG/VR: 43/6.9%, of whom 27.9% discontinued). The most common reasons for treatment discontinuation were adverse events (FCR: 6/30.0%; BR: 11/17.7%; BTKi-based regimens: 60/19.8%; venetoclax-based: 6/8.2%). CONCLUSION: The findings of this study highlight the continued need for tolerable therapies in CLL, with finite therapy offering a better tolerated option for patients who are newly diagnosed or relapsed/refractory to prior treatments.

The effects of musical and linguistic components in recognition of real-world musical excerpts by cochlear implant recipients and normal-hearing adults.

BACKGROUND: Cochlear implants (CI) are effective in transmitting salient features of speech, especially in quiet, but current CI technology is not well suited in transmission of key musical structures (e.g., melody, timbre). It is possible, however, that sung lyrics, which are commonly heard in real-world music may provide acoustical cues that support better music perception. OBJECTIVE: The purpose of this study was to examine how accurately adults who use CIs (n = 87) and those with normal hearing (NH) (n = 17) are able to recognize real-world music excerpts based upon musical and linguistic (lyrics) cues. RESULTS: CI recipients were significantly less accurate than NH listeners on recognition of real-world music with or, in particular, without lyrics; however, CI recipients whose devices transmitted acoustic plus electric stimulation were more accurate than CI recipients reliant upon electric stimulation alone (particularly items without linguistic cues). Recognition by CI recipients improved as a function of linguistic cues. METHODS: Participants were tested on melody recognition of complex melodies (pop, country, & classical styles). Results were analyzed as a function of: hearing status and history, device type (electric only or acoustic plus electric stimulation), musical style, linguistic and musical cues, speech perception scores, cognitive processing, music background, age, and in relation to self-report on listening acuity and enjoyment. Age at time of testing was negatively correlated with recognition performance. CONCLUSIONS: These results have practical implications regarding successful participation of CI users in music-based activities that include recognition and accurate perception of real-world songs (e.g., reminiscence, lyric analysis, & listening for enjoyment).

A review of the incidence of tumor lysis syndrome in patients with chronic lymphocytic leukemia treated with venetoclax and debulking strategies.

We reviewed the literature (January 2010-June 2021) on the effectiveness of debulking strategies before venetoclax initiation in patients with chronic lymphocytic leukemia to reduce tumor burden, downgrade tumor lysis syndrome (TLS) risk, and avoid hospitalization. Low TLS incidence and reduced TLS risk based on tumor burden were reported following debulking in clinical trials. Real-world observational studies reporting debulking regimens recorded no TLS events, and those without debulking strategies had greater TLS incidence. Debulking prior to venetoclax considerably reduces TLS incidence. Further clinical trials and real-world studies may provide additional evidence on effectiveness of debulking in reducing TLS incidence and hospitalization need.

Simulation study of power and sample size for repeated measures with multinomial outcomes: an application to sound direction identification experiments (SDIE).

This study focuses on sample size determination in repeated measures studies with multinomial outcomes from multiple factors. In settings where multiple factors have repeated measures, a single subject could have hundreds of observations. Sample size selection may then refer to the number of subjects, the number of levels within a factor, or the number of repetitions within the level. We simulate multinomial data through a generalized linear mixed model (GLMM) with and without overdispersion, compute the empirical power of detecting group difference for several analytical methods and contrast their performance in group comparison studies with repeated multinomial data. We use four spatial functions to model the spatial correlation structures among observations. We evaluate the factors affecting the power under various scenarios. We also present a dataset typical in hearing studies for sound localization, in which a spatially distributed array of audio loudspeakers plays multiple sounds in order to compare two programming schemes for a hearing aid device.

Temporal stability of music perception and appraisal scores of adult cochlear implant recipients.

BACKGROUND: An extensive body of literature indicates that cochlear implants (CIs) are effective in supporting speech perception of persons with severe to profound hearing losses who do not benefit to any great extent from conventional hearing aids. Adult CI recipients tend to show significant improvement in speech perception within 3 mo following implantation as a result of mere experience. Furthermore, CI recipients continue to show modest improvement as long as 5yr postimplantation. In contrast, data taken from single testing protocols of music perception and appraisal indicate that CIs are less than ideal in transmitting important structural features of music, such as pitch, melody, and timbre. However, there is presently little information documenting changes in music perception or appraisal over extended time as a result of mere experience. PURPOSE: This study examined two basic questions: (1) Do adult CI recipients show significant improvement in perceptual acuity or appraisal of specific music listening tasks when tested in two consecutive years? (2) If there are tasks for which CI recipients show significant improvement with time, are there particular demographic variables that predict those CI recipients most likely to show improvement with extended CI use? RESEARCH DESIGN: A longitudinal cohort study. Implant recipients return annually for visits to the clinic. STUDY SAMPLE: The study included 209 adult cochlear implant recipients with at least 9 mo implant experience before their first year measurement. DATA COLLECTION AND ANALYSIS: Outcomes were measured on the patient's annual visit in two consecutive years. Paired t-tests were used to test for significant improvement from one year to the next. Those variables demonstrating significant improvement were subjected to regression analyses performed to detect the demographic variables useful in predicting said improvement. RESULTS: There were no significant differences in music perception outcomes as a function of type of device or processing strategy used. Only familiar melody recognition (FMR) and recognition of melody excerpts with lyrics (MERT-L) showed significant improvement from one year to the next. After controlling for the baseline value, hearing aid use, months of use, music listening habits after implantation, and formal musical training in elementary school were significant predictors of FMR improvement. Bilateral CI use, formal musical training in high school and beyond, and a measure of sequential cognitive processing were significant predictors of MERT-L improvement. CONCLUSION: These adult CI recipients as a result of mere experience demonstrated fairly consistent music perception and appraisal on measures gathered in two consecutive years. Gains made tend to be modest, and can be associated with characteristics such as use of hearing aids, listening experiences, or bilateral use (in the case of lyrics). These results have implications for counseling of CI recipients with regard to realistic expectations and strategies for enhancing music perception and enjoyment.

Effects of training on recognition of musical instruments presented through cochlear implant simulations.

BACKGROUND: The simulation of the CI (cochlear implant) signal presents a degraded representation of each musical instrument, which makes recognition difficult. PURPOSE: To examine the efficiency and effectiveness of three types of training on recognition of musical instruments as presented through simulations of the sounds transmitted through a CI. RESEARCH DESIGN: Participants were randomly assigned to one of three training conditions: repeated exposure, feedback, and direct instruction. STUDY SAMPLE: Sixty-six adults with normal hearing. INTERVENTION: Each participant completed three training sessions per week, over a five-week time period, in which they listened to the CI simulations of eight different musical instruments. DATA COLLECTION AND ANALYSIS: Analyses on percent of instruments identified correctly showed statistically significant differences between recognition accuracy of the three training conditions (p < .01). RESULTS: those assigned to the direct instruction group showed the greatest improvement over the five-week training period as well as sustained improvement after training. The feedback group achieved the next highest level of recognition accuracy. The repeated exposure group showed modest improvement during the first three-week time period, but no subsequent improvements. CONCLUSIONS: These results indicate that different types of training are differentially effective with regard to improving recognition of musical instruments presented through a degraded signal, which has practical implications for the auditory rehabilitation of persons who use cochlear implants.

Exploring the role of the patient-physician relationship on insulin adherence and clinical outcomes in type 2 diabetes: Insights from the MOSAIc study.

BACKGROUND: The 2-year prospective MOSAIc (Multinational Observational Study assessing Insulin use: understanding the challenges associated with progression of therapy) study is investigating whether patient-, physician-, and health system-related factors affect outcomes in patients with type 2 diabetes (T2D). This baseline subanalysis investigated how aspects of the patient-physician relationship are associated with diabetes-related distress, insulin adherence, and glycemic control. METHODS: Patients with T2D taking insulin for ≥3 months were recruited at primary care and specialty practice sites in 18 countries. Physicians provided usual care. Clinical history and most recent HbA1c values were collected; patients were surveyed regarding their perception of physician interactions, diabetes-related distress level, and insulin adherence. RESULTS: The analysis population comprised 4341 patients. Four (of six) domains showed a significant relationship with total diabetes-related distress (P < 0.01). Poor insulin adherence was associated with greater diabetes-related distress (adjusted odds ratio [aOR] 1.14; 95% confidence interval [CI] 1.06-1.22), higher Discrimination (aOR 1.13; 95% CI 1.02-1.27) and Hurried Communication (aOR 1.35; 95% CI 1.20-1.53) scores, and a lower Explained Results score (aOR 0.86; 95% CI 0.77-0.97). Poor insulin adherence was associated with a 0.43% increase in HbA1c, whereas a 1-unit increase in total diabetes-related distress and Hurried Communication scores was associated with a 0.171% and 0.145% increase in HbA1c, respectively. CONCLUSIONS: Patients distressed about living with T2D, and dissatisfied with aspects of their interactions with physicians, exhibited poor insulin adherence. Perceived physician inattention and lack of engagement (and diabetes-related distress) directly affect insulin adherence and glycemic control.

Impact of Diagnostic Delay and Associated Factors on Clinical Outcomes in a U.S. Inflammatory Bowel Disease Cohort.

BACKGROUND: The impact of diagnostic delay in inflammatory bowel disease, including Crohn's disease (CD) and ulcerative colitis (UC), on disease course remains uncertain. This study examines factors that may influence time to diagnosis and disease outcomes in a U.S. patient cohort. METHODS: We retrospectively collected data on patient characteristics, time to diagnosis, disease phenotype, and complications in 177 patients with inflammatory bowel disease (110 CD and 67 UC) diagnosed at our institution from 2008 to 2015. Factors potentially affecting time to diagnosis were analyzed. Association between disease complications (perianal disease, intestinal strictures, surgery, fistula, abscess, and perforation) and time to diagnosis was tested by multivariable analysis. RESULTS: The median time to diagnosis was longer for patients with CD compared with patients with UC (median 9.5 versus 3.1 months; P < 0.001). The median time from symptom onset to initial physician visit was similar for patients with CD and patients with UC (1 month). However, the median time from symptom onset to specialist evaluation was longer for patients with CD compared with patients with UC: 7 months (interquartile range: 3-23) versus 3 months (interquartile range: 1-8), respectively (P < 0.001). In CD, ileal disease and hematochezia were positively and negatively correlated, respectively, with longer time to diagnosis (P < 0.05). Compared with patients with time to diagnosis ≤4 months, patients with time >26 months had increased overall complications (odds ratio, 8.22; P = 0.007) and intestinal strictures (odds ratio, 8.96; P = 0.012) at time of diagnosis. Such correlation persisted at follow-up. CONCLUSIONS: Time to diagnosis was long in CD. Physician-related delay in diagnosing CD was associated with increased overall complications and intestinal strictures (See Video Abstract, Supplemental Digital Content 1, http://links.lww.com/IBD/B646).

Social economic costs of type 1 diabetes mellitus in pediatric patients in Spain: CHRYSTAL observational study.

AIMS: To estimate the social-economic costs of Type 1 Diabetes Mellitus (T1DM) in patients aged 0-17years in Spain from a social perspective. METHODS: We conducted a cross-sectional observational study in 2014 of 275 T1DM pediatric outpatients distributed across 12 public health centers in Spain. Data on demographic and clinical characteristics, healthcare utilization and informal care were collected from medical records and questionnaires completed by clinicians and patients' caregivers. RESULTS: A valid sample of 249 individuals was analyzed. The average annual cost for a T1DM patient was €27,274. Direct healthcare costs were €4070 and direct non-healthcare cost were €23,204. Informal (familial) care represented 83% of total cost, followed by medical material (8%), outpatient and primary care visits (3.1%) and insulin (2.1%). Direct healthcare cost per patient statistically differed by glycated haemoglobin (HbA1c) level [mean cost €4704 in HbA1c ≥7.5% (≥58mmol/mol) group vs. €3616 in HbA1c<7.5% (<58mmol/mol) group)]; and by the presence or absence of complications and comorbidities (mean cost €5713 in group with complications or comorbidities vs. €3636 in group without complications or comorbidities). CONCLUSIONS: T1DM amongst pediatric patients incurs in considerable societal costs. Informal care represents the largest cost category.

A clinical prognostic prediction of lymph node-negative breast cancer by gene expression profiles.

PURPOSE: To set up a method by use of gene expression data to predict the prognosis of breast cancer patients on the basis of genes as few as possible, but maintaining the accuracy of prediction, we reanalyze the data from van't Veer et al. (Nature 415:530-536, 2002) and van de Vijver et al. (N Engl J Med 347:1999-2009, 2002). METHODS: A three-step method based on re-sampling strategy is employed to select the prognostic genes. And based on these genes, a predictive approach is established. Validation sets are used to testify the predictive power of the prognostic genes. RESULTS: We have discovered 13 genes as the most informative ones to predict the clinical outcomes of breast cancer patients with lymph node-negative. The validation results show the robust performances of these genes. And the results of further analysis illustrate the significant association of the prediction to the time of metastases and overall survival. CONCLUSION: Our predictive approach is useful in prognosis prediction for breast cancer patients with lymph node-negative. The gene markers provide valuable information for the progression of breast cancer and suggest potential target genes for treating the cancer.

Treatment progression in sulfonylurea and dipeptidyl peptidase-4 inhibitor cohorts of type 2 diabetes patients on metformin.

BACKGROUND: Metformin is an oral antidiabetic drug (OAD) widely used as first-line therapy in type 2 diabetes (T2D) treatments. Numerous treatment pathways after metformin failure exist. It is important to understand how treatment choices influence subsequent therapy progressions. This retrospective study compares adherence to, persistence with, and treatment progression in sulfonylurea (SU) and dipeptidyl peptidase-4 (DPP-4) inhibitor patient cohorts with T2D on metformin. METHODS: Using health insurance claims data, matched patient cohorts were created and OAD use was compared in patients with T2D initiating SU or DPP-4 inhibitors (index drugs) since January 1, 2010, to December 31, 2010, with background metformin therapy. Propensity score matching adjusted for possible selection bias. Persistence was measured via Cox regression as days to a ≥60-day gap in index drug possession; adherence was defined as proportion of days covered (PDC) ≥80%. Evolving treatment patterns were traced at 6-month intervals for 24 months following index drug discontinuation. RESULTS: From among 19,621 and 7,484 patients in the SU and DPP-4 inhibitor cohorts, respectively, 6,758 patient pairs were matched. Persistence at 12 months in the SU cohort was 48.0% compared to 52.5% for the DPP-4 inhibitor cohort. PDC adherence (mean [SD]) during the 12-month follow-up period was 63.3 (29.7) for the SU cohort and 65.5 (28.7) for the DPP-4 inhibitor cohort. PDC ≥80% was 40.5% and 43.4% in the SU and DPP-4 inhibitor cohorts, respectively. A higher percentage of patients in the SU cohort remained untreated. Following index drug discontinuation, monotherapy was more common in the SU cohort, while use of two or three OADs was more common in the DPP-4 inhibitor cohort. Insulin therapy initiation was higher in the SU cohort. CONCLUSION: Slightly better adherence and persistence were seen in the DPP-4 inhibitor cohort. Adherence and persistence remain a challenge to many patients; understanding therapy progression will help identify target areas for intervention and improvement.

The Effect of Receiving Treatment Within a Clinical Trial Setting on Survival and Quality of Care Perception in Advanced Stage Non-Small Cell Lung Cancer.

OBJECTIVES: Treatment outcomes of advanced stage (IIIB and IV) non-small cell lung cancer (NSCLC) are poor. In this study, we explore the survival outcomes and the perception of the quality of care delivered in stage IIIB and IV NSCLC patients treated within versus outside a clinical trial. MATERIALS AND METHODS: Data were obtained from the Cancer Care Outcomes Research and Surveillance Consortium (CanCORS). Baseline characteristics according to clinical trial participation were determined. The association between clinical trial enrollment and survival was assessed using a Cox proportional hazard model after adjusting for age, income, primary data collection and research site, comorbidities, self-reported performance status, presence of brain metastasis, stage IIIB versus IV, and cancer histology. RESULTS: Of 815 stage IIIB and IV NSCLC patients, 56 (7%) were enrolled in clinical trials. Median survival for the patients treated within versus outside a clinical trial was 20.5 versus 16.7 months, respectively (P=0.21). Using a multivariate survival model, clinical trial enrollment did not correlate with longer survival (P=0.81). Comparing patients according to clinical trial enrollment, patients treated within a clinical trial setting perceived a better overall quality of care (P<0.01). CONCLUSIONS: Management of stage IIIB and IV NSCLC patients within a clinical trial setting conveyed a perception of superior care that did not translate into survival benefit. These findings suggest that providing cancer care within a clinical trial should not imply a survival benefit when counseling stage IIIB and IV NSCLC patients about entering clinical trials.

Cross-National Variation in Glycemic Control and Diabetes-Related Distress Among East Asian Patients Using Insulin: Results from the MOSAIc Study.

INTRODUCTION: Guidelines recommend insulin progression for patients with type 2 diabetes (T2D) with inadequate glycemic control. The Multinational Observational Study Assessing Insulin use (MOSAIc [ClinicalTrials.gov identifier, NCT01400971]) study is a 2-year observational study, investigating factors that influence insulin progression in T2D patients. In this first of two reports, we describe baseline clinical and psychosocial characteristics of Chinese, Japanese, and South Korean patients who participated in MOSAIc. Insulin treatment, factors affecting progression, and outcomes will be reported separately. METHODS: Patients with T2D using insulin for ≥3 months were eligible. Baseline demographic, clinical, and psychosocial data were collected from patients. Quality of life instruments, including the Diabetes Distress Scale (DDS), were used to assess patient's concerns about disease management, support, and emotional burden. The association between the DDS and the selected covariates was also assessed. RESULTS: A total of 373 patients in China, 157 in Japan, and 141 in South Korea were enrolled from July 2011 to July 2013. Mean ± standard deviation duration (years) of T2D differed across countries (China 11.4 ± 7.5; Japan 13.8 ± 8.7; South Korea 15.7 ± 8.8; P < 0.0001). Japanese patients used more noninsulin anti-hyperglycemic agents than did Chinese or South Korean patients (P < 0.0001). Exclusive use of basal insulin was most common in Japan and South Korea compared with China, whereas approximately 66.8% of Chinese patients used mixed insulin. Covariates associated with the DDS were younger age [P = 0.044 (Japan)], higher incidence of monthly hypoglycemia [P = 0.036 [China]; P = 0.021 (South Korea)], and male gender [P = 0.037 (South Korea)]. CONCLUSIONS: There were significant differences amongst East Asian patients with T2D treated with insulin, including in quality of life scores. Results from the MOSAIc longitudinal analyses will further investigate trends of insulin intensification and barriers to insulin progression. FUNDING: Eli Lilly and Company.

Incidence, prevalence, and trend analysis of the use of insulin delivery systems in the United States (2005 to 2011).

OBJECTIVE: Estimate the incidence and prevalence rates and assess overall trends among patients with diabetes using insulin vial/syringe and pens over time. METHODS: A retrospective study was conducted using the Truven MarketScan database from 1 July 2004 to 31 December 2011. The database contained medical and pharmacy claims of >20 million US residents insured by commercial health plans. Patients with diabetes who utilized vial/syringe and pens were included. Incidence rate was defined as the proportion of patients initiating a new insulin type for the first time with vial/syringe and pens within each year from 2005 to 2011 among the total number of patients initiating that particular insulin type for the first time. Prevalence rate was defined as the proportion of patients using vial/syringe or pens among patients with diabetes using that insulin type within each year from 2005 to 2011. A linear trend over time was assessed by Cochran-Armitage Trend tests and Generalized Estimating Equations. RESULTS: Incidence of patients initiating vial/syringe decreased from 2005 to 2011 (basal analog [90.5% to 31.3%]; mealtime analog [67.6% to 37.1%]), while patients initiating pens increased (basal analog [9.5% to 68.7%]; mealtime analog [32.4% to 62.9%]). There was a significant trend over time indicating increased usage of pens relative to vial/syringe; the number of pen users increased (all p < 0.0001 except mealtime human). Prevalence of patients using vial/syringe decreased from 2005 to 2011 (basal analog [93.8% to 41.2%]; mealtime analog [71.0% to 50.6%]), while patients using pens increased (basal analog [6.2% to 58.8%]; mealtime analog [29.0% to 49.4%]). From 2005 to 2011, patients were more likely to use pens than vial/syringe (all p < 0.0001 except for human mixtures and mealtime human). CONCLUSIONS: The incidence and prevalence of patients using the traditional vial/syringe decreased over time, while the use of pens increased. Some patient populations may be under-represented, limiting generalizability of results.

Associations Between Glycemic Control, Depressed Mood, Clinical Depression, and Diabetes Distress Before and After Insulin Initiation: An Exploratory, Post Hoc Analysis.

INTRODUCTION: Although depression is often associated with poor glycemic control in patients with type 2 diabetes mellitus (T2DM), this observation has been inconsistent. This exploratory, post hoc analysis investigated associations between depression parameters and glycemic control using data from a 24-month, prospective, observational, non-interventional study evaluating glycemic response following insulin initiation for T2DM. METHODS: We analyzed data from a 24-month, prospective, observational study that evaluated glycemic response in patients with T2DM who initiated insulin therapy (N = 985) in 5 European countries. Secondary measures included patient-reported diagnosis of depression at baseline, severity of depressed/anxious mood (EuroQol (EQ)-5D item) and diabetes-related distress (Psychological Distress domain of the Diabetes Health Profile, DHP-18). The latter two measures were assessed at baseline and 5 time points throughout the study. Glycemic control was measured by glycated hemoglobin (HbA1c) at these same time points. Analyses employed t tests to assess the unadjusted baseline difference in HbA1c between patients with and without the respective depression parameter. The potential effect of demographic and clinical confounding variables was controlled through a linear model structure. Patient HbA1c levels were analyzed by presence/absence of a history of diagnosed depression, depressed mood, and diabetes-related distress. RESULTS: Patients with higher depression parameters or distress at baseline had significantly higher rates of microvascular complications at baseline. Patients with a history of diagnosed depression or high diabetes-related distress had higher HbA1c than patients without. HbA1c of patients with or without depressed mood was not significantly different at baseline. The proportion of patients with depressed mood declined after insulin initiation, whereas the proportion of patients with high diabetes-related distress did not significantly change. HbA1c improved following insulin initiation, regardless of presence/absence of studied depression/distress parameters at baseline. CONCLUSION: History of diagnosed depression, diabetes-related distress, and depressed mood were associated with a higher rate of microvascular complications. Diagnosed depression and diabetes-related distress also showed higher HbA1c at baseline when insulin was initiated. Insulin therapy improved glycemic control, while preexisting depressed mood declined and diabetes-related distress remained unchanged.

Majorization Minimization by Coordinate Descent for Concave Penalized Generalized Linear Models.

Recent studies have demonstrated theoretical attractiveness of a class of concave penalties in variable selection, including the smoothly clipped absolute deviation and minimax concave penalties. The computation of the concave penalized solutions in high-dimensional models, however, is a difficult task. We propose a majorization minimization by coordinate descent (MMCD) algorithm for computing the concave penalized solutions in generalized linear models. In contrast to the existing algorithms that use local quadratic or local linear approximation to the penalty function, the MMCD seeks to majorize the negative log-likelihood by a quadratic loss, but does not use any approximation to the penalty. This strategy makes it possible to avoid the computation of a scaling factor in each update of the solutions, which improves the efficiency of coordinate descent. Under certain regularity conditions, we establish theoretical convergence property of the MMCD. We implement this algorithm for a penalized logistic regression model using the SCAD and MCP penalties. Simulation studies and a data example demonstrate that the MMCD works sufficiently fast for the penalized logistic regression in high-dimensional settings where the number of covariates is much larger than the sample size.

The cross-validated AUC for MCP-logistic regression with high-dimensional data.

We propose a cross-validated area under the receiving operator characteristic (ROC) curve (CV-AUC) criterion for tuning parameter selection for penalized methods in sparse, high-dimensional logistic regression models. We use this criterion in combination with the minimax concave penalty (MCP) method for variable selection. The CV-AUC criterion is specifically designed for optimizing the classification performance for binary outcome data. To implement the proposed approach, we derive an efficient coordinate descent algorithm to compute the MCP-logistic regression solution surface. Simulation studies are conducted to evaluate the finite sample performance of the proposed method and its comparison with the existing methods including the Akaike information criterion (AIC), Bayesian information criterion (BIC) or Extended BIC (EBIC). The model selected based on the CV-AUC criterion tends to have a larger predictive AUC and smaller classification error than those with tuning parameters selected using the AIC, BIC or EBIC. We illustrate the application of the MCP-logistic regression with the CV-AUC criterion on three microarray datasets from the studies that attempt to identify genes related to cancers. Our simulation studies and data examples demonstrate that the CV-AUC is an attractive method for tuning parameter selection for penalized methods in high-dimensional logistic regression models.