Medical student characteristics predictive of intent for rural practice.

INTRODUCTION: The shortage of physicians in rural areas of the USA has led medical schools to focus on recruiting and selecting students who will choose to work in non-urban settings. The purpose of this study was to examine the effect of student characteristics previously correlated to choosing rural practice (ie being older, being male, being raised in a rural community, having a spouse or significant other who was raised in a rural community, having a spouse or significant other whose parents live in a rural community, having parents with high school education or less, and graduating from a smaller college) on osteopathic medical students' intent to practice in a rural setting. This study also examined the correlation between personality type and intent for rural practice using the Myers Briggs Type Indicator (MBTI). Finally, this study examined factors that would increase interest in practicing in a rural setting, such as financial assistance and students' opinions about physicians who choose to practice in rural areas. METHODS: The study participants were students in a new osteopathic medical school with an enrollment of 225. Cross-sectional survey data were collected on background characteristics and intent for rural practice. Retrospective data were collected from MBTI assessments previously completed as part of routine career planning educational sessions. Data were analyzed using descriptive statistics such as frequencies, percentages and inferential statistics such as χ² and logistic regression. RESULTS: A total of 141 students participated, a 63% response rate. Factors associated with intent to practice rural medicine included being raised in a rural area for more than half of one's life (p<0.05) and having a spouse or significant other who had lived in a rural area (p<0.05). Age, sex, race, and size of undergraduate college were not associated with intent to practice rural medicine. Students categorized as Extraverted based on the MBTI were more likely to have intent for rural practice even when other factors were controlled. Students reported that financial incentives and wage guarantees may increase interest in rural practice. CONCLUSION: The results of this study support past research showing that medical students with a rural background and with spouses or significant others having a rural background are more likely to have intent for rural practice. This study also found that students' personality types may be correlated with intent to practice in a rural area. In order to provide physicians who will serve the population living in rural areas of the USA, it is imperative that medical schools select students who are most likely to practice in a non-urban setting. Financial incentives are important to students, suggesting that programs such as loan forgiveness may be useful in recruiting students to rural practice. Medical students may benefit from career counseling utilizing the MBTI to facilitate an understanding of personality type and how it may impact their preference for rural practice.

A simulation study of predictive ability measures in a survival model II: explained randomness and predictive accuracy.

Several R(2) -type measures have been proposed to evaluate the predictive ability of a survival model. In Part I, we classified the measures into four categories and studied the measures in the explained variation category. In this paper, we study the remaining measures in a similar fashion, discussing their strengths and shortcomings. Simulation studies are used to examine the performance of the measures with respect to the criteria we set out in Part I. Our simulation studies showed that among the measures studied in this paper, the measures proposed by Kent and O'Quigley ρ(W)(2) (and its approximation ρ(W,A)(2)) and Schemper and Kaider R(SK)(2) perform better with respect to our criteria. However, our investigations showed that ρ(W)(2) is adversely affected by the distribution of covariate and the presence of influential observations. The results show that the other measures perform poorly, primarily because they are affected either by the degree of censoring or the follow-up period.

Estimating the crude probability of death due to cancer and other causes using relative survival models.

Relative survival is used extensively in population-based cancer studies to measure patient survival correcting for causes of death not related to the disease of interest. An advantage of relative survival is that it provides a measure of mortality associated with a particular disease, without the need for information on cause of death. Relative survival provides a measure of net mortality, i.e. the probability of death due to cancer in the absence of other causes. This is a useful measure, but it is also of interest to measure crude mortality, i.e. the probability of death due to cancer in the presence of other causes. A previous approach to estimate the crude probability of death in population-based cancer studies used life table methods, but we show how the estimates can be obtained after fitting a relative survival model. We adopt flexible parametric models for relative survival, which use restricted cubic splines for the baseline cumulative excess hazard and for any time-dependent effects. We illustrate the approach using an example of men diagnosed with prostate cancer in England and Wales showing the differences in net and crude survival for different ages.

How do multi-stage, multi-arm trials compare to the traditional two-arm parallel group design--a reanalysis of 4 trials.

BACKGROUND: To speed up the evaluation of new therapies, the multi-arm, multi-stage trial design was suggested previously by the authors. METHODS: In this paper, we evaluate the performance of the two-stage, multi-arm design using four cancer trials conducted at the MRC CTU. The performance of the design at fictitious interim analyses is assessed using a conditional bootstrap approach. RESULTS: Two main aims are addressed: the error rate of correctly carrying on/stopping the trial at an interim analysis as well as quantifying the gains in terms of resources by employing this design. Furthermore, we make suggestions for the best timing of this interim analysis. CONCLUSION: Multi-arm, multi-stage trials are an effective way of speeding up the therapy evaluation process. The design performs well in terms of the type I and II error rates.

Prognostic factors for renal cell carcinoma.

Renal cell carcinoma is a relatively uncommon tumour with a widely varying prognosis depending on several tumour and clinical factors. This review discusses these factors and critically appraises their value both as individual markers and when they are incorporated into scoring systems/models or algorithms. Disease stage (assessed pathologically and/or clinically) and performance status have the strongest evidence as helpful individual prognostic markers but a better discrimination is obtained by combining these and adding in various other indices. Prospective validation of such integrated prognostic models will be essential.

An approach to estimating prognosis using fractional polynomials in metastatic renal carcinoma.

We present a prognostic model for metastatic renal cell carcinoma based on fractional polynomials. We retrospectively analysed 425 metastatic renal cell carcinoma patients treated with subcutaneous recombinant cytokine-based home therapies in consecutive trials. In our approach, we categorised a continuous prognostic index produced by the multivariable fractional polynomial (MFP) algorithm, using a strategy in which continuous predictors are kept continuous. The MFP algorithm selected five prognostic factors as significant at the 5% level in a multivariable model: lymph node metastases, liver metastases, bone metastases, age, C-reactive protein and neutrophils. The MFP model allowed us to divide patients into four risk groups achieving median overall survivals of 38 months (low risk), 23 months (low intermediate risk), 15 months (high intermediate risk) and 5.6 months (high risk). Our approach, based on categorising a continuous prognostic index produced by the MFP algorithm, allowed more flexibility in the determination of risk groups than traditional approaches.

Prediction of patient-specific risk of early preterm delivery using maternal history and sonographic measurement of cervical length: a population-based prospective study.

OBJECTIVE: To develop a model for calculating the patient-specific risk of spontaneous early preterm delivery by combining maternal factors and the transvaginal sonographic measurement of cervical length at 22 + 0 to 24 + 6 weeks, and to compare the detection rate of this method to that achieved from screening by cervical length or maternal characteristics alone. METHODS: This was a population-based prospective multicenter study involving 40,995 unselected women with singleton pregnancies attending for routine hospital antenatal care in London, UK. Complete follow-up was obtained from 39,284 (95.8%) cases. The main outcomes were detection rate, false-positive rate and accuracy of predicting spontaneous delivery before 32 weeks' gestation. RESULTS: Spontaneous delivery before 32 weeks occurred in 235 (0.6%) cases. The detection rate of screening for early preterm delivery, at a fixed false-positive rate of 10%, was 38% for maternal factors, 55% for cervical length and 69% for combined testing. There was good agreement between the model estimates and the observed probabilities of preterm delivery. CONCLUSIONS: This study provides a model that can give an accurate patient-specific risk of preterm delivery. The detection rate of screening by a combination of maternal factors and the measurement of cervical length was substantially higher than that of screening by each method alone.

Evaluation of sample size and power for multi-arm survival trials allowing for non-uniform accrual, non-proportional hazards, loss to follow-up and cross-over.

We present a general framework for sample size calculation in survival studies based on comparing two or more survival distributions using any one of a class of tests including the logrank test. Incorporated within this framework are the possible presence of non-uniform staggered patient entry, non-proportional hazards, loss to follow-up and treatment changes including cross-over between treatment arms. The framework is very general in nature and is based on using piecewise exponential distributions to model the survival distributions. We illustrate the use of the approach and explore its validity using simulation studies. These studies have shown that not adjusting for loss to follow-up, non-proportional hazards or cross-over can lead to significant alterations in power or equivalently, a marked effect on sample size. The approach has been implemented in the freely available program ART (for Stata). Our investigations suggest that ART is the first software to allow incorporation of all these elements. Further extensions to the methodology such as non-local alternatives for the logrank test are also considered.

Building multivariable regression models with continuous covariates in clinical epidemiology--with an emphasis on fractional polynomials.

OBJECTIVES: In fitting regression models, data analysts must often choose a model based on several candidate predictor variables which may influence the outcome. Most analysts either assume a linear relationship for continuous predictors, or categorize them and postulate step functions. By contrast, we propose to model possible non-linearity in the relationship between the outcome and several continuous predictors by estimating smooth functions of the predictors. We aim to demonstrate that a structured approach based on fractional polynomials can give a broadly satisfactory practical solution to the problem of simultaneously identifying a subset of 'important' predictors and determining the functional relationship for continuous predictors. METHODS: We discuss the background, and motivate and describe the multivariable fractional polynomial (MFP) approach to model selection from data which include continuous and categorical predictors. We compare our results with those from other approaches in examples. We present a small simulation study to compare the functional form of the relationship obtained by fitting fractional polynomials and splines to a single predictor variable. RESULTS: We illustrate the advantages of the MFP approach over standard techniques of model construction in two real example datasets analyzed with logistic and Cox regression models, respectively. In the simulation study, fractional polynomial models had lower mean square error and more realistic behaviour than comparable spline models. CONCLUSIONS: In many practical situations, the MFP approach can satisfy the aim of finding models that fit the data well and also are simple, interpretable and potentially transportable to other settings.

Is treatment with interferon-alpha effective in all patients with metastatic renal carcinoma? A new approach to the investigation of interactions.

The first analysis of the MRC RE01 trial in metastatic renal carcinoma identified a 28% reduction in the hazard of death for patients treated with interferon-alpha compared with medroxyprogesterone acetate (MPA). No subgroup was identified in which treatment with interferon-alpha was more or less effective than MPA. We used a new approach based on fractional polynomials to investigate the updated data from this trial for the possible interaction of treatment with prognostic factors. In the spirit of hypothesis generation, we considered 10 possible prognostic variables, of which white cell count (WCC) was found to influence the effectiveness of interferon treatment. In patients treated with MPA, there was no prognostic effect of WCC, whereas, in patients treated with interferon, the risk of dying increased significantly with WCC level. We defined subgroups of patients based on WCC levels and estimated a hazard ratio of 0.53 in favour of interferon in patients with WCC <6.5 x 10(9), whereas for patients with WCC >10 x 10(9) the risk appears to be similar between the treatment groups, or even slightly raised in the interferon group. Since our results are derived from flexible statistical models, they may be interpreted as a new hypothesis and require validation in independent data.

Stability of multivariable fractional polynomial models with selection of variables and transformations: a bootstrap investigation.

Sauerbrei and Royston have recently described an algorithm, based on fractional polynomials, for the simultaneous selection of variables and of suitable transformations for continuous predictors in a multivariable regression setting. They illustrated the approach by analyses of two breast cancer data sets. Here we extend their work by considering how to assess possible instability in such multivariable fractional polynomial models. We first apply the algorithm repeatedly in many bootstrap replicates. We then use log-linear models to investigate dependencies among the inclusion fractions for each predictor and among the simplified classes of fractional polynomial function chosen in the bootstrap samples. To further evaluate the results, we define measures of instability based on a decomposition of the variability of the bootstrap-selected functions in relation to a reference function from the original model. For each data set we are able to identify large, reasonably stable subsets of the bootstrap replications in which the functional forms of the predictors appear fairly stable. Despite the considerable flexibility of the family of fractional polynomials and the consequent risk of overfitting when several variables are considered, we conclude that the multivariable selection algorithm can find stable models.

Metastatic renal carcinoma comprehensive prognostic system.

The purpose of the study was to identify a comprehensive prognostic system of pretreatment clinical parameters in 425 patients (pts) with metastatic renal-cell carcinoma treated with different subcutaneous (s.c.) recombinant cytokine-based home therapies in consecutive trials. Treatment consisted of (A) s.c. interferon-alpha 2a (INF-alpha), s.c. interleukin-2 (IL-2) (n=102 pts), (B) s.c. IFN-alpha 2a, s.c. IL-2, and i.v. 5-fluorouracil (5-FU) (n=235 pts) or (C) s.c. IFN-alpha 2a, s.c. IL-2, and i.v. 5-FU combined with p.o. 13-cis-retinoic acid (13cRA) (n=88 pts). Kaplan-Meier survival analysis, log-rank statistics, and Cox regression analysis were employed to identify risk factors and to create a multiple risk factor model. The following pretreatment risk factors were identified by univariate analysis: (1) three and more metastatic sites, (2) presence of liver, lymph node or bone metastases, (3) neutrophil count > or = 6500 cells microl(-1), (4) serum lactate dehydrogenase level (LDH) > or = 220 U l(-1), and (5) serum C-reactive protein level (CRP) > or = 11 mg l(-1). Cox regression analysis with forward stepwise variable selection identified neutrophil count as the major prognostic factor (hazard ratio=1.9, P<0.001), while serum levels of LDH and CRP, time between diagnosis of tumour and onset of metastatic disease, number of metastatic sites, and bone metastases were significant but somewhat less important prognostic variables within the multiple risk factor model (hazard ratio < or = 1.5). Patients were assigned to one of the three risk groups according to cumulative risk defined as the sum of simplified risk s.c.ores for six pretreatment variables. Low-, intermediate-, and high-risk patients achieved a median overall survival of 32+ months (95% CI 24, 43; 5-year survival of 27%), 18+ months (95% CI 15, 20; 5-year survival of 11%), and 8+ months (95% CI 6, 10; 5-year survival of 5%), respectively. These prognostic categories are helpful both in individual patient care and in the assessment of patients entering prospective clinical trials.

The effect of stopping smoking on cervical Langerhans' cells and lymphocytes.

OBJECTIVE: To investigate the effects of stopping smoking on cervical Langerhans' cells and lymphocytes. DESIGN: Prospective intervention study. SETTING: A large family planning clinic in central London. POPULATION: Women volunteers prepared to attempt to give up smoking for six months. Their most recent cervical smear showed no abnormality greater than mild dyskaryosis. METHODS: The women were seen at three-month intervals for six months. Reduction in smoking was assessed by self-reporting and validated by salivary cotinine concentrations. Colposcopy and a biopsy of a normal area were performed at the first and last visits. Any area of abnormality was also biopsied at the final visit. Langerhans' cells and lymphocytes were counted. MAIN OUTCOME MEASURES: Proportional changes in counts of Langerhans' cells and lymphocytes with reduction in smoking. RESULTS: Reduction in smoking by 20 to 40 cigarettes per day was significantly associated with a reduction of between 6% and 16% in counts of Langerhans cells, CD8 and total lymphocytes. Heavy smoking was significantly associated (P = 0.02) with an increased chance of persistent human papillomavirus infection. The presence of candida was associated with significantly higher counts of between 41% and 47% in total lymphocytes and CD8 lymphocytes. In contrast, the presence of anaerobic vaginosis was associated with significantly lower counts of between 16% and 30% in Langerhans cells, CD4 and CD8 lymphocytes. CONCLUSIONS: This large intervention study has demonstrated a clear relationship between reduction in smoking and changes in cervical immune cell counts. Future studies need to take into account cytokine interactions, which recent studies suggest may be significant in the immune response to both human papillomavirus and cervical intraepithelial neoplasia and the ever-increasing complexity of the cell-mediated immune system of the cervix.

Reference values of fetal peak systolic blood flow velocity in the middle cerebral artery at 19-40 weeks of gestation.

OBJECTIVES: The objectives of this prospective study were (i) to establish new reference values of peak systolic blood flow velocity measurement in the fetal middle cerebral artery (MCA-PSV) following validated methodological guidelines and (ii) to develop a method to calculate Z-scores of MCA-PSV. PATIENTS AND METHODS: Cross-sectional data were obtained from 331 pregnant women between 19 and 40 weeks' gestation. Reference ranges for MCA-PSV were constructed and for each measurement linear regression models were fitted separately to the mean and standard deviations (SD) as a function of gestational age. An application to calculate Z-scores was developed. A comparison was made between the reference ranges produced in our study and those of a previous one. RESULTS: A new chart, table of centiles and regression equations of MCA-PSV are presented. Comparison of our reference ranges with ones produced in a previous study showed similar 5th centile values. However, the values for the 50th and 95th centiles between 19 and 28 gestational weeks were lower in our study. CONCLUSIONS: We have constructed reference ranges for MCA-PSV which, because they are derived from a larger number of examinations in the 15-20-week period and because the methodological flaws of the previously published study have been eliminated, we consider to be more accurate and therefore more useful for clinical practice.

Goodness-of-fit statistics for age-specific reference intervals.

The age-specific reference interval is a commonly used screening tool in medicine. It involves estimation of extreme quantile curves (such as the 5th and 95th centiles) of a reference distribution of clinically normal individuals. It is crucial that models used to estimate such intervals fit the data extremely well. However, few procedures to assess goodness-of-fit have been proposed in the literature, and even fewer have been evaluated systematically. Here we consider procedures based on the distribution of the Z-scores (standardized residuals) from a model and on Pearson chi(2) statistics for observed and expected counts in groups defined by age and the estimated reference centile curves. Two of the procedures (Q and grid tests) are mainly inferential, whereas the third (permutation bands and B-tests) is essentially graphical. We obtain approximations to the null distributions of several relevant test statistics and examine their size and power for a range of models based on real data sets. We recommend Q-tests in all situations where Z-scores are available since they are general, simple to calculate and usually have the highest power among the three classes of test considered. For the cases considered the grid tests are always inferior to the Q- and B- tests.

A useful monotonic non-linear model with applications in medicine and epidemiology.

In medicine and epidemiology monotonic curves are important as models for relations which prior knowledge or scientific reasoning dictate should increase or decrease consistently with the predictor value. An example is the monotonically increasing relation between cigarette consumption and the risk of coronary heart disease. In this paper I propose a new class of monotonic non-linear models which generalizes the well-known power and exponential transformations of a covariate. The models are cousins of the Gompertz family of growth curves and include non-sigmoid and asymmetric sigmoid curves. I explore their properties and illustrate their usefulness in three substantial medical and epidemiological data sets.

A strategy for modelling the effect of a continuous covariate in medicine and epidemiology.

Low-dimensional parametric models are well understood, straightforward to communicate to other workers, have very smooth curves and may easily be checked for consistency with background scientific knowledge or understanding. They should therefore be ideal tools with which to represent smooth relationships between a continuous predictor and an outcome variable in medicine and epidemiology. Unfortunately, a seriously restricted set of such models is used routinely in practical data analysis - typically, linear, quadratic or occasionally cubic polynomials, or sometimes a power or logarithmic transformation of a covariate. Since their flexibility is limited, it is not surprising that the fit of such models is often poor. Royston and Altman's recent work on fractional polynomials has extended the range of available functions. It is clearly crucial that the chosen final model fits the data well. Achieving a good fit with minimal restriction on the functional form has been the motivation behind the major recent research effort on non-parametric curve-fitting techniques. Here I propose that one such model, a (possibly over-fitted) cubic smoothing spline, may be used to define a suitable reference curve against which the fit of a parametric model may be checked. I suggest a significance test for the purpose and examine its type I error and power in a small simulation study. Several families of parametric models, including some with sigmoid curves, are considered. Their suitability in fitting regression relationships found in several real data sets is investigated. With all the example data sets, a simple parametric model can be found which fits the data approximately as well as a cubic smoothing spline, but without the latter's tendency towards artefacts in the fitted curve.

Choice of scale for cubic smoothing spline models in medical applications.

The determination of the functional form of the relationship between an outcome variable and one or more continuous covariates is an important aspect of the modelling of medical data. For correct interpretation of the data it is essential that the functional form be specified at least approximately correctly. I show that for given model complexity, logarithmic transformation of a covariate can greatly improve the fit of one of the most useful and convenient non-parametric regression models, the cubic smoothing spline. A mathematical rationale for the idea is given. I propose a diagnostic for deciding initially whether a log transformation is needed. The method is illustrated using several medical data sets. No special software other than that used for fitting the spline models is needed.

What do we mean by validating a prognostic model?

Prognostic models are used in medicine for investigating patient outcome in relation to patient and disease characteristics. Such models do not always work well in practice, so it is widely recommended that they need to be validated. The idea of validating a prognostic model is generally taken to mean establishing that it works satisfactorily for patients other than those from whose data it was derived. In this paper we examine what is meant by validation and review why it is necessary. We consider how to validate a model and suggest that it is desirable to consider two rather different aspects - statistical and clinical validity - and examine some general approaches to validation. We illustrate the issues using several case studies.