Sample size re-estimation for clinical trials with longitudinal negative binomial counts including time trends

In some diseases, such as multiple sclerosis, lesion counts obtained from magnetic resonance imaging (MRI) are used as markers of disease progression. This leads to longitudinal, and typically overdispersed, count data outcomes in clinical trials. Models for such data invariably include a number of nuisance parameters, which can be difficult to specify at the planning stage, leading to considerable uncertainty in sample size specification. Consequently, blinded sample size re-estimation procedures are used, allowing for an adjustment of the sample size within an ongoing trial by estimating relevant nuisance parameters at an interim point, without compromising trial integrity. To date, the methods available for re-estimation have required an assumption that the mean count is time-constant within patients. We propose a new modeling approach that maintains the advantages of established procedures but allows for general underlying and treatment-specific time trends in the mean response. A simulation study is conducted to assess the effectiveness of blinded sample size re-estimation methods over fixed designs. Sample sizes attained through blinded sample size re-estimation procedures are shown to maintain the desired study power without inflating the Type I error rate and the procedure is demonstrated on MRI data from a recent study in multiple sclerosis.     

Nonproportional hazards and unobserved heterogeneity in clustered survival data: When can we tell the difference?

Multivariate survival data are frequently encountered in biomedical applications in the form of clustered failures (or recurrent events data). A popular way of analyzing such data is by using shared frailty models, which assume that the proportional hazards assumption holds conditional on an unobserved cluster-specific random effect. Such models are often incorporated in more complicated joint models in survival analysis. If the random effect distribution has finite expectation, then the conditional proportional hazards assumption does not carry over to the marginal models. It has been shown that, for univariate data, this makes it impossible to distinguish between the presence of unobserved heterogeneity (eg, due to missing covariates) and marginal nonproportional hazards. We show that time-dependent covariate effects may falsely appear as evidence in favor of a frailty model also in the case of clustered failures or recurrent events data, when the cluster size or number of recurrent events is small. When true unobserved heterogeneity is present, the presence of nonproportional hazards leads to overestimating the frailty effect. We show that this phenomenon is somewhat mitigated as the cluster size grows. We carry out a simulation study to assess the behavior of test statistics and estimators for frailty models in such contexts. The gamma, inverse Gaussian, and positive stable shared frailty models are contrasted using a novel software implementation for estimating semiparametric shared frailty models. Two main questions are addressed in the contexts of clustered failures and recurrent events: whether covariates with a time-dependent effect may appear as indication of unobserved heterogeneity and whether the additional presence of unobserved heterogeneity can be detected in this case. Finally, the practical implications are illustrated in a real-world data analysis example.     

The Geriatrician: The Frontline Specialist in the Treatment of COVID-19 Patients

On February 20, 2020, a man living in the north of Italy was admitted to the emergency room with an atypical pneumonia that later proved to be COVID-19. This was the trigger of one of the most serious clusters of COVID-19 in the world, outside of China. Despite aggressive restraint and inhibition efforts, COVID-19 continues to increase, and the total number of infected patients in Italy is growing daily. After 6 weeks, the total number of patients reached 128,948 cases (April 5, 2020), with the higher case-fatality rate (15,887 deaths) dominated by old and very old patients.This sudden health emergency severely challenged the Italian Health System, in particular acute care hospitals and intensive care units. In 1 hospital, geriatric observation units were created, the experience of which can be extremely useful for European countries, the United States, and all countries that in the coming days will face a similar situation.     

The Influence of Dysphagia on Nutritional and Frailty Status among Community-Dwelling Older Adults

Malnutrition is a core symptom of the frailty cycle in older adults. The purpose of this study was to investigate whether dysphagia influences nutrition or frailty status in community-dwelling older adults. The study participants were 320 Japanese community-dwelling older adults aged ≥65 years. All participants completed a questionnaire survey that included items on age, sex, family structure, self-rated health, nutritional and frailty status, and swallowing function. Nutritional status was categorized as malnourished, at risk of malnutrition, and well-nourished based on the Mini Nutrition Assessment-Short Form. The participants were then classified into a malnutrition (malnourished/at risk) or a well-nourished group (well-nourished). Frailty was assessed using the Cardiovascular Health Study criteria. The participants were then divided into a frailty (frail/pre-frail) or a non-frailty group (robust). Dysphagia was screened using the 10-item Eating Assessment Tool. Multiple logistic regression analysis was conducted to determine whether dysphagia was associated with nutritional or frailty status. The results revealed that dysphagia influenced both nutrition (odds ratio [OR]: 4.0; 95% confidence interval [CI]: 1.9–8.2) and frailty status (OR: 2.3; 95% CI: 1.0–5.2); therefore, the swallowing function would be an important factor for community-dwelling older adults on frailty prevention programs.     

Role of Cachexia and Fragility in the Patient Candidate for Cardiac Surgery

Frailty is the major expression of accelerated aging and describes a decreased resistance to stressors, and consequently an increased vulnerability to additional diseases in elderly people. The vascular aging related to frail phenotype reflects the high susceptibility for cardiovascular diseases and negative postoperative outcomes after cardiac surgery. Sarcopenia can be considered a biological substrate of physical frailty. Malnutrition and physical inactivity play a key role in the pathogenesis of sarcopenia. We searched on Medline (PubMed) and Scopus for relevant literature published over the last 10 years and analyzed the strong correlation between frailty, sarcopenia and cardiovascular diseases in elderly patient. In our opinion, a right food intake and moderate intensity resistance exercise are mandatory in order to better prepare patients undergoing cardiac operation.     

Concurrent and Predictive Validity of FRAIL-NH in Hospitalized Older Persons: An Exploratory Study

ObjectivesThe FRAIL-NH was originally developed for frailty assessment of nursing home (NH) residents. We aimed to compare concurrent, predictive, and known-groups validity between FRAIL-NH and FRAIL, using the Frailty Index (FI) as gold standard reference. We also examined for ceiling effect of both measures in the detection of severe frailty.DesignA secondary analysis of a prospective cohort study.Setting & ParticipantsOlder adults (mean age 89.4 years) hospitalized for an acute medical illness in a 1300-bed tertiary hospital.MeasurementsBaseline data on demographics, comorbidities, severity of illness, functional status, and cognitive status were gathered. We also captured outcomes of mortality, length of stay (LOS), institutionalization, and functional decline. For concurrent validity, we compared areas under the operating characteristic curves (AUCs) for both measures against the FI. For predictive validity, univariate analyses and multiple logistic regression were used to compare both measures against the adverse outcomes of interest. For known-groups validity, we compared both measures against comorbidities and functional status via 1-way analysis of variance, and dementia diagnosis via independent t test. Box plots were also derived to investigate for possible ceiling effect.ResultsBoth measures had good concurrent validity (both AUC > 0.8 and P < .001), with FRAIL-NH detecting more frailty cases (79.5% vs 50.0%). Although FRAIL-frail was superior for in-hospital mortality [6.7% vs 1.0%, P = .031, odds ratio (OR) 9.29, 95% confidence interval (CI) 1.09-79.20, P < .042] and LOS (10 vs 8 days, P = .043), FRAIL-NH-frail better predicted mortality (OR 6.62, 95% CI 1.91-22.94, P = .003) and institutionalization (OR 6.03, 95% CI 2.01-18.09, P = .001) up to 12 months postenrollment. Known-groups validity was good for both measures with FRAIL-NH yielding greater F values for functional status and dementia. Lastly, box plots revealed a ceiling effect for FRAIL in the severely frail group.Conclusions and ImplicationsThis exploratory study highlights the potential for expanding the role of FRAIL-NH beyond NH to acute care settings. Contrasted to FRAIL, FRAIL-NH had better overall validity with less ceiling effect in discrimination of severe frailty.     

Transitions Between Frailty States and Its Predictors in a Cohort of Community-Dwelling Spaniards

ObjectivesStudy the frequency and determinants of frailty transitions in a community-dwelling older population.DesignPopulation-based prospective longitudinal study [The Toledo Study of Healthy Ageing (TSHA)].Setting and Participants1748 community-dwelling individuals aged >65 years living in Toledo, a Spanish province.MethodsFrailty was measured with the Fried phenotype. Logistic models were used to assess the associations of sociodemographic, clinical, life-habits, functional, physical performance, and analytical variables with frailty transitions (losing robustness, transitioning from prefrailty to robustness, and from prefrailty to frailty) over a median of 5.2 years.ResultsMean age on enrolment was 75 years, and 55.8% were females. At baseline, 10.3% were frail and 43.1% prefrail. At follow-up, 35.8% of the frail individuals recovered to a prefrail and 15.1% to a robust state. In addition, 43.7% of the prefrail participants became robust, but 14.5% developed frailty. Of those robust at baseline, 32.9% became prefrail and 4.2% frail. In multivariate logistic models, chair-stands had a predictive role in all transitions studied: linearly in keeping robustness and with a floor effect (5 stands) in transitions from prefrailty to robustness and (inversely) from prefrailty to frailty. More depressive symptoms were associated with unfavorable transitions. Not declaring the amount of alcohol drunk and low grip strength were associated with loss of robustness. Hearing and cognitive impairment, low physical activity and smoking with transitioning from prefrailty to frailty. Autonomy for instrumental activities of daily living and uricemia were associated with transitions between robustness and prefrailty in both directions. Increasing body mass index in the range of moderate to severe obesity hampered regaining robustness.Conclusions and ImplicationsSpontaneous improvement of frailty measured with the Fried phenotype is frequent, mainly to prefrailty. Most of the variables associated with transitions are modifiable and suggest research topics and interventions to reduce frailty in clinical and social care settings.     

Frailty in Nursing Homes—A Prospective Study Comparing the FRAIL-NH and the Clinical Frailty Scale

ObjectivesFrailty is common in nursing home (NH) residents, but its prevalence in German institutions is unknown. Valid and easy-to-use screening tools are needed to identify frail residents. We used the FRAIL-NH scale and the Clinical Frailty Scale (CFS) to (1) obtain the prevalence of frailty, (2) investigate the agreement between both instruments, and (3) evaluate their predictive validity for adverse health events in German NH residents.DesignProspective cohort study.Setting and participantsGerman NH residents (n = 246, age 84 ± 8 years, 67% female).MethodsFrailty status was categorized according to FRAIL-NH (nonfrail, frail, most frail) and CFS (not frail, mild to moderately frail, severely frail). Agreement between instruments was examined by Spearman correlation, an area under the receiver operating characteristic curve (AUC) with 95% CI, and sensitivity and specificity using the “most frail” category of FRAIL-NH as reference standard. Adverse health events (death, hospital admissions, falls) were recorded for 12 months, and multivariate cox and logistic regression models calculated.ResultsAccording to FRAIL-NH, 71.1% were most frail, 26.4% frail, and 2.5% nonfrail. According to CFS, 66.3% were severely frail, 26.8% mild to moderately frail, and 6.9% not frail. Both scales correlated significantly (r = 0.78; R² = 60%). The AUC was 0.92 (95% CI 0.88-0.96). Using a CFS cutoff of 7 points, sensitivity was 0.90 and specificity 0.92. The frailest groups according to both instruments had an increased risk of death [FRAIL-NH hazard ratio (HR) 2.19, 95% CI 1.21-3.99; CFS HR 2.56, 95% CI 1.43-4.58] and hospital admission [FRAIL-NH odds ratio (OR) 1.95, 95% CI 1.06-3.58; CFS OR 1.79, 95% CI 1.01-3.20] compared to less frail residents. The FRAIL-NH predicted recurrent faller status (OR 2.57, 95% CI 1.23-5.39).Conclusions and implicationsFrailty is highly prevalent in German NH residents. Both instruments show good agreement despite different approaches and are able to predict adverse health outcomes. Based on our findings and because of its simple administration, CFS may be an alternative to FRAIL-NH for assessing frailty in NHs.