Anika Buchholz

Comparison of procedures to assess non-linear and time-varying effects in multivariable models for survival data

The focus of many medical applications is to model the impact of several factors on time to an event. A standard approach for such analyses is the Cox proportional hazards model. It assumes that the factors act linearly on the log hazard function (linearity assumption) and that their effects are constant over time (proportional hazards (PH) assumption). Variable selection is often required to specify a more parsimonious model aiming to include only variables with an influence on the outcome. As follow-up increases the effect of a variable often gets weaker, which means that it varies in time. However, spurious time-varying effects may also be introduced by mismodelling other parts of the multivariable model, such as omission of an important covariate or an incorrect functional form of a continuous covariate. These issues interact. To check whether the effect of a variable varies in time several tests for non-PH have been proposed. However, they are not sufficient to derive a model, as appropriate modelling of the shape of time-varying effects is required. In three examples we will compare five recently published strategies to assess whether and how the effects of covariates from a multivariable model vary in time. For practical use we will give some recommendations.

On properties of predictors derived with a two-step bootstrap model averaging approach-A simulation study in the linear regression model

In many applications of model selection there is a large number of explanatory variables and thus a large set of candidate models. Selecting one single model for further inference ignores model selection uncertainty. Often several models fit the data equally well. However, these models may differ in terms of the variables included and might lead to different predictions. To account for model selection uncertainty, model averaging procedures have been proposed. Recently, an extended two-step bootstrap model averaging approach has been proposed. The first step of this approach is a screening step. It aims to eliminate variables with negligible effect on the outcome. In the second step the remaining variables are considered in bootstrap model averaging. A large simulation study is performed to compare the MSE and coverage rate of models derived with bootstrap model averaging, the full model, backward elimination using Akaike and Bayes information criterion and the model with the highest selection probability in bootstrap samples. In a data example, these approaches are also compared with Bayesian model averaging. Finally, some recommendations for the development of predictive models are given.

On stability issues in deriving multivariable regression models.

In many areas of science where empirical data are analyzed, a task is often to identify important variables with influence on an outcome. Most often this is done by using a variable selection strategy in the context of a multivariable regression model. Using a study on ozone effects in children (n = 496, 24 covariates), we will discuss aspects relevant for deriving a suitable model. With an emphasis on model stability, we will explore and illustrate differences between predictive models and explanatory models, the key role of stopping criteria, and the value of bootstrap resampling (with and without replacement). Bootstrap resampling will be used to assess variable selection stability, to derive a predictor that incorporates model uncertainty, check for influential points, and visualize the variable selection process. For the latter two tasks we adapt and extend recent approaches, such as stability paths, to serve our purposes. Based on earlier experiences and on results from the example, we will argue for simpler models and that predictions are usually very similar, irrespective of the selection method used. Important differences exist for the corresponding variances, and the model uncertainty concept helps to protect against serious underestimation of the variance of a predictor-derived data dependently. Results of stability investigations illustrate severe difficulties in the task of deriving a suitable explanatory model. It seems possible to identify a small number of variables with an important and probably true influence on the outcome, but too often several variables are included whose selection may be a result of chance or may depend on a small number of observations.

Shared decision-making in antihypertensive therapy: a cluster randomised controlled trial

BackgroundHypertension is one of the key factors causing cardiovascular diseases. A substantial proportion of treated hypertensive patients do not reach recommended target blood pressure values. Shared decision making (SDM) is to enhance the active role of patients. As until now there exists little information on the effects of SDM training in antihypertensive therapy, we tested the effect of an SDM training programme for general practitioners (GPs). Our hypotheses are that this SDM training (1) enhances the participation of patients and (2) leads to an enhanced decrease in blood pressure (BP) values, compared to patients receiving usual care without prior SDM training for GPs.MethodsThe study was conducted as a cluster randomised controlled trial (cRCT) with GP practices in Southwest Germany. Each GP practice included patients with treated but uncontrolled hypertension and/or with relevant comorbidity. After baseline assessment (T0) GP practices were randomly allocated into an intervention and a control arm. GPs of the intervention group took part in the SDM training. GPs of the control group treated their patients as usual. The intervention was blinded to the patients. Primary endpoints on patient level were (1) change of patients’ perceived participation (SDM-Q-9) and (2) change of systolic BP (24h-mean). Secondary endpoints were changes of (1) diastolic BP (24h-mean), (2) patients’ knowledge about hypertension, (3) adherence (MARS-D), and (4) cardiovascular risk score (CVR).ResultsIn total 1357 patients from 36 general practices were screened for blood pressure control by ambulatory blood pressure monitoring (ABPM). Thereof 1120 patients remained in the study because of uncontrolled (but treated) hypertension and/or a relevant comorbidity. At T0 the intervention group involved 17 GP practices with 552 patients and the control group 19 GP practices with 568 patients. The effectiveness analysis could not demonstrate a significant or relevant effect of the SDM training on any of the endpoints.ConclusionThe study hypothesis that the SDM training enhanced patients’ perceived participation and lowered their BP could not be confirmed. Further research is needed to examine the impact of patient participation on the treatment of hypertension in primary care.Trial registrationGerman Clinical Trials Register (DRKS): DRKS00000125

Implementation of shared decision making by physician training to optimise hypertension treatment. Study protocol of a cluster-RCT

BackgroundHypertension is one of the key factors causing cardiovascular diseases which make up the most frequent cause of death in industrialised nations. However about 60% of hypertensive patients in Germany treated with antihypertensives do not reach the recommended target blood pressure. The involvement of patients in medical decision making fulfils not only an ethical imperative but, furthermore, has the potential of higher treatment success. One concept to enhance the active role of patients is shared decision making. Until now there exists little information on the effects of shared decision making trainings for general practitioners on patient participation and on lowering blood pressure in hypertensive patients.Methods/DesignIn a cluster-randomised controlled trial 1800 patients receiving antihypertensives will be screened with 24 h ambulatory blood pressure monitoring in their general practitioners’ practices. Only patients who have not reached their blood pressure target (approximately 1200) will remain in the study (T1 – T3). General practitioners of the intervention group will take part in a shared decision making-training after baseline assessment (T0). General practitioners of the control group will treat their patients as usual. Primary endpoints are change of systolic blood pressure and change of patients’ perceived participation. Secondary endpoints are changes of diastolic blood pressure, knowledge, medical adherence and cardiovascular risk. Data analysis will be performed with mixed effects models.DiscussionThe hypothesis underlying this study is that shared decision making, realised by a shared decision making training for general practitioners, activates patients, facilitates patients’ empowerment and contributes to a better hypertension control. This study is the first one that tests this hypothesis with a (cluster-) randomised trial and a large sample size.Trial registrationWHO International Clinical Trials: http://apps.who.int/trialsearch/Trial.aspx?TrialID=DRKS00000125

Improving the Prognostic Ability through Better Use of Standard Clinical Data - The Nottingham Prognostic Index as an Example

Background Prognostic factors and prognostic models play a key role in medical research and patient management. The Nottingham Prognostic Index (NPI) is a well-established prognostic classification scheme for patients with breast cancer. In a very simple way, it combines the information from tumor size, lymph node stage and tumor grade. For the resulting index cutpoints are proposed to classify it into three to six groups with different prognosis. As not all prognostic information from the three and other standard factors is used, we will consider improvement of the prognostic ability using suitable analysis approaches. Methods and Findings Reanalyzing overall survival data of 1560 patients from a clinical database by using multivariable fractional polynomials and further modern statistical methods we illustrate suitable multivariable modelling and methods to derive and assess the prognostic ability of an index. Using a REMARK type profile we summarize relevant steps of the analysis. Adding the information from hormonal receptor status and using the full information from the three NPI components, specifically concerning the number of positive lymph nodes, an extended NPI with improved prognostic ability is derived. Conclusions The prognostic ability of even one of the best established prognostic index in medicine can be improved by using suitable statistical methodology to extract the full information from standard clinical data. This extended version of the NPI can serve as a benchmark to assess the added value of new information, ranging from a new single clinical marker to a derived index from omics data. An established benchmark would also help to harmonize the statistical analyses of such studies and protect against the propagation of many false promises concerning the prognostic value of new measurements. Statistical methods used are generally available and can be used for similar analyses in other diseases.

DECADE-pilot: decision aid, action planning, and follow-up support for patients to reduce the 10-year risk of cardiovascular diseases—a protocol of a randomized controlled pilot trial

BackgroundA healthy lifestyle can reduce cardiovascular risk (CVR) and prevent premature death. Usually most patients at increased CVR have difficulties implementing the necessary health behavior changes, such as smoking cessation, increasing of physical activity, healthy diet, stress reduction, etc. In this pilot study, a new intervention (DECADE) that includes a cardiovascular risk calculation, evidence-based decision aids, action planning, and follow-up support for patients to reduce their 10-year risk of cardiovascular diseases will be tested in primary care. The objectives of this trial are to test (1) the feasibility of the study design in preparation for the main trail including (2) the usability and acceptance of DECADE, and (3) initial data to ascertain that changes can be observed in these patients.MethodsThis randomized controlled pilot trial will generate initial data on the potential effects of DECADE on patients’ self-evaluated activity and behavior change as well as on clinical outcomes such as blood pressure, cholesterol, body mass index (BMI), HbA1C, and CVR score. In the qualitative part of the study, we will analyze data collected in semi-structured interviews with participating general practitioners (GP) and in patient questionnaires.DiscussionThe outcomes of this pilot study will indicate whether DECADE is a promising intervention in the domain of patient-centered prevention of cardiovascular diseases (CVD) and whether a larger multi-center randomized controlled trial is feasible.Trial registrationGerman Clinical Trials Register (DRKS), DRKS00010584

Treatment of Primary Breast Cancer at the Surgical Unit of the Charité 1984-1998

Background: We have analyzed the patient population of one clinic (Charité) over a period of 15 years. Besides the changes in the technical facilities and therapeutical guidelines during these years, this period also reflects the changes in the health system attributable to the reunification of East and West Germany. Until now only few analyses for breast cancer patients from the German speaking area have been reported. Patients and Methods: All 2,062 patients undergoing surgical treatment for breast cancer between 1984 and 1998 were documented and followed up until 2007. The analysis included 1,560 patients with a primary breast cancer who fulfilled certain inclusion criteria. The treatment strategies applied to this population are presented in 3 time periods (1984-1990, 1991-1993, and 1994-1998). The effects of prognostic factors on overall survival were investigated using univariate analyses. Results: The percentage of pT1 tumors changed from 50.7% in the first period to 63.1% in the third period. The percentage of node-negative patients hardly changed with time (on average 61.6%). However, the percentage of patients with less than 10 assessed nodes decreased from 48.4% to 6.7% and 2.5% for the 3 periods, respectively. Therapeutic strategies changed drastically. Survival rate increased substantially, most likely due to improved therapeutic strategies, but also for other reasons not considered in the analysis.