Perceived helpfulness of the individual components of a behavioural weight loss program: results from the Hopkins POWER Trial

Background

Behavioural weight loss programs are effective first‐line treatments for obesity and are recommended by the US Preventive Services Task Force. Gaining an understanding of intervention components that are found helpful by different demographic groups can improve tailoring of weight loss programs. This paper examined the perceived helpfulness of different weight loss program components.

Methods

Participants (n = 236) from the active intervention conditions of the Practice‐based Opportunities for Weight Reduction (POWER) Hopkins Trial rated the helpfulness of 15 different components of a multicomponent behavioural weight loss program at 24‐month follow‐up. These ratings were examined in relation to demographic variables, treatment arm and weight loss success.

Results

The components most frequently identified as helpful were individual telephone sessions (88%), tracking weight online (81%) and coach review of tracking (81%). The component least frequently rated as helpful was the primary care providers' general involvement (50%). Groups such as older adults, Blacks and those with lower education levels more frequently reported intervention components as helpful compared with their counterparts.

Discussion

Weight loss coaching delivered telephonically with web support was well received. Findings support the use of remote behavioural interventions for a wide variety of individuals.

Patient-reported outcomes in the practice-based opportunities for weight reduction (POWER) trial

PURPOSE

To evaluate effects of two behavioral weight-loss interventions (in-person, remote) on health-related quality of life (HRQOL) compared to a control intervention.

METHODS

Four hundred and fifty-one obese US adults with at least one cardiovascular risk factor completed five measures of HRQOL and depression: MOS SF-12 physical component summary (PCS) and mental component summary; EuroQoL-5 dimensions single index and visual analog scale; PHQ-8 depression symptoms; and PSQI sleep quality scores at baseline and 6 and 24 months after randomization. Change in each outcome was analyzed using outcome-specific mixed-effects models controlling for participant demographic characteristics.

RESULTS

PCS-12 scores over 24 months improved more among participants in the in-person active intervention arm than among control arm participants (P < 0.05, ES = 0.21); there were no other statistically significant treatment arm differences in HRQOL change. Greater weight loss was associated with improvements in most outcomes (P < 0.05 to < 0.0001).

CONCLUSIONS

Participants in the in-person active intervention improved more in physical function HRQOL than participants in the control arm did. Greater weight loss during the study was associated with greater improvement in all PRO except for sleep quality, suggesting that weight loss is a key factor in improving HRQOL.

Geographic prevalence and multilevel determination of community-level factors associated with herpes simplex virus type 2 infection in Chennai, India

Herpes simplex virus type 2 (HSV-2) is one of the most prevalent sexually transmitted infections, and it increases the risk of transmission of human immunodeficiency virus type 1 at least twofold. Individual-level factors are insufficient to explain geographic and population variation in HSV-2, suggesting the need to identify ecologic factors. The authors sought to determine the geographic prevalence and community-level factors associated with HSV-2 after controlling for individual-level factors among slums in Chennai, India. From March to June 2001, participants aged 18-40 years voluntarily completed a survey and were tested for HSV-2. Community characteristics were assessed through interviews with key informants and other secondary data sources. Multilevel nonlinear analysis was conducted. Eighty-five percent of eligible persons completed the survey; of these, 98% underwent HSV-2 testing, producing a final sample of 1,275. Participants were of Tamil ethnicity, were predominantly female and married, and were on average 30 years old. Fifteen percent were infected with HSV-2, and there was significant variation in HSV-2 prevalence among communities. After controlling for individual-level factors, the authors identified community-level factors, including socioeconomic status and the presence of injection drug users, that were independently associated with HSV-2 and explained 11% of the variance in prevalence. Future studies are needed to test mechanisms through which these community-level factors may be operating.

State-level adjusted ESRD incident rates: use of observed vs model-predicted category-specific rates

Because of differences in case-mix across states, state-level case-mix-adjusted end-stage renal disease (ESRD) incident rates are reported in each United States Renal Data System Annual Data Report to make the across-state comparisons valid. The adjusted rates were estimated by the direct adjustment method, a widely used method for adjusted event rate calculation, based on observed category-specific ESRD incident rates in each state (called the observation-based method). However, when some adjusting categories in a state are small, the adjusted rate and the standard error for this state as estimated by this method may be inaccurate. This report proposes a model-based method that can overcome the disadvantages of the observation-based method and can be extended to continuous adjusting variables. National ESRD incident data and national population data from 1990 to 1999 were used. State-level adjusted ESRD incident rates were estimated by both the observation- and the model-based methods. For the model-based method, a Poisson regression model was used to estimate category-specific ESRD incident rates. For large-population states, both observation- and model-based methods produced similar estimates for adjusted ESRD incident rates. For small-population states, however, the observation-based method produced year-to-year estimates of adjusted ESRD incident rates that varied considerably and also had very large standard errors. In contrast, the model-based method produced stable estimates. The model-based method can overcome the disadvantages of the observation-based method for estimating state-level adjusted ESRD incident rates, especially for small states.

National Institutes of Health Consensus Development Conference Statement: adjuvant therapy for breast cancer, November 1-3, 2000

OBJECTIVE

Our goal was to provide health-care providers, patients, and the general public with an assessment of currently available data regarding the use of adjuvant therapy for breast cancer.

PARTICIPANTS

The participants included a non-Federal, non-advocate, 14-member panel representing the fields of oncology, radiology, surgery, pathology, statistics, public health, and health policy as well as patient representatives. In addition, 30 experts in medical oncology, radiation oncology, biostatistics, epidemiology, surgical oncology, and clinical trials presented data to the panel and to a conference audience of 1000.

EVIDENCE

The literature was searched with the use of MEDLINE(TM) for January 1995 through July 2000, and an extensive bibliography of 2230 references was provided to the panel. Experts prepared abstracts for their conference presentations with relevant citations from the literature. Evidence from randomized clinical trials and evidence from prospective studies were given precedence over clinical anecdotal experience.

CONSENSUS PROCESS

The panel, answering predefined questions, developed its conclusions based on the evidence presented in open forum and the scientific literature. The panel composed a draft statement, which was read in its entirety and circulated to the experts and the audience for comment. Thereafter, the panel resolved conflicting recommendations and released a revised statement at the end of the conference. The panel finalized the revisions within a few weeks after the conference. The draft statement was made available on the World Wide Web immediately after its release at the conference and was updated with the panel's final revisions. The statement is available at http://consensus.nih.gov.

CONCLUSIONS

The panel concludes that decisions regarding adjuvant hormonal therapy should be based on the presence of hormone receptor protein in tumor tissues. Adjuvant hormonal therapy should be offered only to women whose tumors express hormone receptor protein. Because adjuvant polychemotherapy improves survival, it should be recommended to the majority of women with localized breast cancer regardless of lymph node, menopausal, or hormone receptor status. The inclusion of anthracyclines in adjuvant chemotherapy regimens produces a small but statistically significant improvement in survival over non-anthracycline-containing regimens. Available data are currently inconclusive regarding the use of taxanes in adjuvant treatment of lymph node-positive breast cancer. The use of adjuvant dose-intensive chemotherapy regimens in high-risk breast cancer and of taxanes in lymph node-negative breast cancer should be restricted to randomized trials. Ongoing studies evaluating these treatment strategies should be supported to determine if such strategies have a role in adjuvant treatment. Studies to date have included few patients older than 70 years. There is a critical need for trials to evaluate the role of adjuvant chemotherapy in these women. There is evidence that women with a high risk of locoregional tumor recurrence after mastectomy benefit from postoperative radiotherapy. This high-risk group includes women with four or more positive lymph nodes or an advanced primary cancer. Currently, the role of postmastectomy radiotherapy for patients with one to three positive lymph nodes remains uncertain and should be tested in a randomized controlled trial. Individual patients differ in the importance they place on the risks and benefits of adjuvant treatments. Quality of life needs to be evaluated in selected randomized clinical trials to examine the impact of the major acute and long-term side effects of adjuvant treatments, particularly premature menopause, weight gain, mild memory loss, and fatigue. Methods to support shared decision-making between patients and their physicians have been successful in trials; they need to be tailored for diverse populations and should be tested for broader dissemination.

Triple-goal estimates for disease mapping

Maps of regional morbidity and mortality rates play an important role in assessing environmental equity. They provide effective tools for identifying areas with potentially elevated risk, determining spatial trend, and formulating and validating aetiological hypotheses about disease. Bayes and empirical Bayes methods produce stable small-area rate estimates that retain geographic and demographic resolution. The beauty of the Bayesian approach lies in its ability to structure complicated models, inferential goals and analyses. Three inferential goals are relevant to disease mapping and risk assessment: (i) computing accurate estimates of disease rates in small geographic areas; (ii) estimating the distribution of disease rates over the region; (iii) ranking the disease rates so that environmental investigation can be prioritized. No single set of estimates can simultaneously optimize these three goals, and Shen and Louis propose a set of estimates that perform well on all three goals. These are optimal for estimating the distribution of rates and for ranking, and maintain a high accuracy in estimating area-specific rates. However, the Shen/Louis method is sensitive to choice of priors. To address this issue we introduce a robustified version of the method based on a smoothed non-parametric estimate of the prior. We evaluate the performance of this method through a simulation study, and illustrate it using a data set of county-specific lung cancer rates in Ohio.

A linear mixed-effects model for multivariate censored data

We apply a linear mixed-effects model to multivariate failure time data. Computation of the regression parameters involves the Buckley-James method in an iterated Monte Carlo expectation-maximization algorithm, wherein the Monte Carlo E-step is implemented using the Metropolis-Hastings algorithm. From simulation studies, this approach compares favorably with the marginal independence approach, especially when there is a strong within-cluster correlation.

Effect of calcium supplementation on serum cholesterol and blood pressure. A randomized, double-blind, placebo-controlled, clinical trial

OBJECTIVE

To test the effect of daily supplemental calcium on serum total and high-density lipoprotein cholesterol (HDL-C) levels and blood pressure in adults.

DESIGN

Randomized, double-blind, placebo-controlled clinical trial; adjunct study to a trial of calcium and colon cell proliferation in patients with sporadic adenoma.

SETTING

Outpatient clinic.

PATIENTS

A total of 193 men and women, aged 30 to 74 years.

INTERVENTION

Treatment with 1.0 and 2.0 g/d of elemental calcium vs placebo over a 4-month period for cholesterol determinations and 6 months for blood pressure.

MAIN OUTCOME MEASURES

Serum total cholesterol and HDL-C levels, systolic and diastolic blood pressure.

RESULTS

Because there were no apparent differences in responses between the 1.0-g and 2.0-g calcium groups, their data were combined and compared with those of the placebo group. Among all participants, the mean total cholesterol level dropped 0.07 mmol/L (2.9 mg/dL) (1.3%) (P = .43) more, and the mean HDL-C level dropped 0.01 mmol/L (0.4 mg/dL) (1.1%) (P = .71) less in the calcium group than in the placebo group. Among participants without a history of hypercholesterolemia, the mean total cholesterol level dropped 0.18 mmol/L (6.8 mg/dL) (3.3%) (P = .10) and the HDL-C level dropped 0.02 mmol/L (0.6 mg/dL) (1.5%) (P = .61) more in the calcium group than in the placebo group. Among all participants, there was no apparent change in blood pressure until 6 months, when the mean systolic blood pressure dropped 0.8 mm Hg (0.6%) (P = .85) and the mean diastolic blood pressure dropped 0.4 mm Hg (0.5%) (P = .80) more in the calcium group than in the placebo group.

CONCLUSIONS

There were no substantial or statistically significant effects of calcium supplementation on total cholesterol or HDL-C levels or on blood pressure. There was a suggestion (not statistically significant) of a 0.07 to 0.18 mmol/L (3-7 mg/dL) or 2% to 4% drop in the total cholesterol level, a finding similar to that reported in other studies, which indicates the need for further study.

Innovations in bayes and empirical bayes methods: estimating parameters, populations and ranks

By formalizing the relation among components and 'borrowing information' among them, Bayes and empirical Bayes methods can produce more valid, efficient and informative statistical evaluations than those based on traditional methods. In addition, Bayesian structuring of complicated models and goals guides development of appropriate statistical approaches and generates summaries which properly account for sampling and modelling uncertainty. Computing innovations enable implementation of complex and relevant models, thereby substantially increasing the role of Bayes/empirical Bayes methods in important statistical assessments. Policy-relevant statistical assessments involve synthesis of information from a set of related components such as medical clinics, geographic regions or research studies. Typical assessments include inference for individual parameters, synthesis over the collection of components (for example, the parameter histogram) and comparisons among parameters (for example, ranks). The relative importance of these goals depends on the context. Bayesian structuring provides a guide to valid inference. For example, while posterior means are the 'obvious' and optimal estimates for individual components under squared error loss, their empirical distribution function (EDF) is underdispersed and never valid for estimating the EDF of the true, underlying parameters. Effective histogram estimates result from optimizing a loss function based in a distance between the histogram and its estimate. Similarly, ranking observed data usually produces poor estimates and ranking posterior means can be inappropriate. Effective estimates should be based on a loss function that caters directly to ranks. Using examples of 'borrowing information', shrinkage and the variance/bias trade-off we motivate Bayes and empirical Bayes analysis. Then, we outline the formal approach and discuss 'triple-goal' estimates with values that when ranked produce optimal ranks, for which the EDF is an optimal estimate of the parameter EDF and such that the values themselves are effective estimates of co-ordinate-specific parameters. We use basic models and data analysis examples to highlight the conceptual and structural issues.

Innovations in bayes and empirical bayes methods: estimating parameters, populations and ranks

By formalizing the relation among components and 'borrowing information' among them, Bayes and empirical Bayes methods can produce more valid, efficient and informative statistical evaluations than those based on traditional methods. In addition, Bayesian structuring of complicated models and goals guides development of appropriate statistical approaches and generates summaries which properly account for sampling and modelling uncertainty. Computing innovations enable implementation of complex and relevant models, thereby substantially increasing the role of Bayes/empirical Bayes methods in important statistical assessments. Policy-relevant statistical assessments involve synthesis of information from a set of related components such as medical clinics, geographic regions or research studies. Typical assessments include inference for individual parameters, synthesis over the collection of components (for example, the parameter histogram) and comparisons among parameters (for example, ranks). The relative importance of these goals depends on the context. Bayesian structuring provides a guide to valid inference. For example, while posterior means are the 'obvious' and optimal estimates for individual components under squared error loss, their empirical distribution function (EDF) is underdispersed and never valid for estimating the EDF of the true, underlying parameters. Effective histogram estimates result from optimizing a loss function based in a distance between the histogram and its estimate. Similarly, ranking observed data usually produces poor estimates and ranking posterior means can be inappropriate. Effective estimates should be based on a loss function that caters directly to ranks. Using examples of 'borrowing information', shrinkage and the variance/bias trade-off we motivate Bayes and empirical Bayes analysis. Then, we outline the formal approach and discuss 'triple-goal' estimates with values that when ranked produce optimal ranks, for which the EDF is an optimal estimate of the parameter EDF and such that the values themselves are effective estimates of co-ordinate-specific parameters. We use basic models and data analysis examples to highlight the conceptual and structural issues.

Expressing estimators of expected quality adjusted survival as functions of Nelson-Aalen estimators

Quality adjusted survival has been increasingly advocated in clinical trials to be assessed as a synthesis of survival and quality of life. We investigate nonparametric estimation of its expectation for a general multistate process with incomplete follow-up data. Upon establishing a representation of expected quality adjusted survival through marginal distributions of a set of defined events, we propose two estimators for expected quality adjusted survival. Expressed as functions of Nelson-Aalen estimators, the two estimators are strongly consistent and asymptotically normal. We derive their asymptotic variances and propose sample-based variance estimates, along with evaluation of asymptotic relative efficiency. Monte Carlo studies show that these estimation procedures perform well for practical sample sizes. We illustrate the methods using data from a national, multicenter AIDS clinical trial.

Environmental justice and statistical summaries of differences in exposure distributions

Recent regulatory action requires the assessment of environmental justice (equitable protection from the burdens of environmental hazards across sociodemographic subpopulations) in the siting of hazardous waste sites, and prioritization of environmental remediation efforts. Assessments of environmental justice require linking exposure, demographic, and health data. The geographic nature of the data makes the use of geographic information systems attractive for environmental justice assessments. Typical geographic assessments compare the composition of 'exposed' populations, while typical statistical assessments focus on differences in health outcomes between population subgroups, possibly adjusted for exposure. We outline an alternate approach based on summarized differences between exposure distributions within each population subgroup. We illustrate how such summaries provide a tool for site evaluation (e.g., defining exposure inequities resulting from locating a new potential hazard at any of a number of possible sites). In addition, we describe summaries, based on dose-response relationships, to describe risk differences imposed by the observed exposure differences. Reported toxic emissions from Allegheny County, Pennsylvania illustrate the approach.

Colorectal adenomatous and hyperplastic polyps: smoking and N-acetyltransferase 2 polymorphisms

Arylamine N-acetyltransferase 2 (NAT2) is involved in both the detoxification and bioactivation of carcinogenic arylamines and other mutagens. This enzyme is polymorphic, and the fast and slow phenotypes are thought to be risk factors for colon and bladder cancer, respectively. Here, we report on a case-control study of adenomatous and hyperplastic polyps, with particular attention to tobacco smoking, a known risk factor for adenomas, and polymorphisms of NAT2. All participants underwent complete colonoscopy and were subsequently divided into case and control groups on the basis of pathology. Cases were diagnosed with confirmed adenomas (n = 527) or hyperplastic polyps (n = 200); controls (n = 633) had no history of colonic neoplasia and no polyps at colonoscopy. NAT2 genotype was determined using an oligonucleotide ligation assay and fast, intermediate, or slow phenotype imputed. Multivariate-adjusted odds ratios (ORs) and 95% confidence intervals were computed using logistic regression adjusting for age, sex, nonsteroidal anti-inflammatory drug use, and hormone replacement therapy use. Smoking was associated with an increased risk of adenomas [current versus never smoking OR = 2.0 (95% confidence interval, 1.4-2.9)] and hyperplastic polyps [current versus never smoking OR = 4.1 (2.6-6.5)]. NAT2 status among adenomatous polyp patients and hyperplastic polyp patients, respectively, showed ORs of 1.1 (0.8-1.4) and 1.2 (0.8-1.6; intermediate versus slow) and 1.1 (0.6-1.9) and 0.9 (0.4-1.9; fast versus slow). There were no differences in risk when adenoma patients were stratified on multiplicity, size, or histopathological subtype of polyps. Never-smokers showed no variation in risk across acetylator status for either species of polyp, whereas current smokers showed ORs of 2.0 (1.2-3.2) and 2.3 (1.4-3.9) for adenomas and 3.9 (2.1-7.1) and 4.9 (2.6-9.4) for hyperplastic polyps for slow and intermediate/fast NAT2, respectively, compared with slow-NAT2 never-smokers. Risks of both multiple [OR = 4.3 (2.1-8.8)] and large [OR = 3.8 (1.9-7.5)] adenomas were somewhat elevated in current smokers with an intermediate/fast phenotype compared with smokers with a slow NAT2 phenotype, but the interaction was not statistically significant. Risk of hyperplastic polyps and adenomatous polyps is strongly related to smoking. There is little suggestion of interaction between NAT2 status and smoking and no relationship with NAT2 genotype alone.

A meta-analysis of the effects of dietary protein restriction on the rate of decline in renal function

Dietary protein restriction has been reported to delay the need for renal replacement therapy in clinical trials and meta-analyses. However, less clear is what effect dietary protein has on the rate of decline in renal function. We pooled the results of 13 randomized controlled trials (n = 1,919 patients) and found that dietary protein restriction reduced the rate of decline in estimated glomerular filtration rate by only 0.53 mL/min/yr (95% confidence interval [CI], 0.08 to 0.98 mL/min/yr). We also used weighted regression analysis to determine the reasons for the differences in the results of these 13 randomized trials along with 11 other nonrandomized controlled trials (n = 2,248 patients). The effect of dietary protein restriction (glomerular filtration rate decline in treatment minus control) was substantially less in randomized versus nonrandomized trials (regression coefficient, -5.2 mL/min/yr; 95% CI, -7.8 to -2.5 mL/min/yr; P < 0.05) and relatively greater among diabetic versus nondiabetic patients (5.4 mL/min/yr; 95% CI, 0.3 to 10.5 mL/min/yr; P < 0.05), while there was a trend toward a greater effect with each additional year of follow-up (2.1 mL/min/yr; 95% CI, -0.05 to 4.2 mL/min/yr; P = NS). However, the number of diabetic patients studied was small and the duration of follow-up was short in most trials. No other patient or study characteristics altered the effect of dietary protein restriction on the rate of decline in renal function. Thus, although dietary protein restriction retards the rate of renal function decline, the relatively weak magnitude of this effect suggests that better therapies are needed to slow the rate of renal disease progression.

A randomized, placebo-controlled trial of the safety and efficacy of oral ganciclovir for prophylaxis of cytomegalovirus disease in HIV-infected individuals. Terry Beirn Community Programs for Clinical Research on AIDS

OBJECTIVE

Evaluate safety and efficacy of oral ganciclovir (GCV) for preventing cytomegalovirus (CMV) disease in HIV-infected persons at high risk for CMV disease.

DESIGN

Double-blind, placebo-controlled, randomized clinical trial in primary care clinics and private practice offices specializing in the care of people with HIV. Interventions were oral GCV (1000 mg three times/day) or placebo. Protocol amendment allowed switch to open-label oral GCV. Main outcome measures were confirmed CMV retinal or gastrointestinal mucosal disease, and death. The study enrolled 994 people co-infected with CMV and HIV, with at least one CD4 count recorded < 100 x 10(6) cells/l.

RESULTS

At study completion (15 months median follow-up), CMV event rates in the oral GCV and control groups were 13.1 and 14.6 per 100 person years, respectively, a hazard ratio (HR) of 0.92 [95% confidence interval (CI), 0.65-1.27; P = 0.6]. At protocol amendment event rates were 12.7 and 15.0, respectively (HR, 0.85; 95% CI, 0.56-1.30; P = 0.45). At study completion, event rates for death were 26.6 and 32.0 (HR, 0.84; P = 0.09), and at protocol amendment were 18.9 and 19.6 (HR, 0.95; P = 0.78), respectively. At protocol amendment for the CMV endpoint, the oral GCV treatment effect was associated with baseline use of didanosine (ddI). For patients taking ddI at randomization, HR was 7.48 (P = 0.02). For patients not taking ddI, HR was 0.62 (P = 0.04). These HR were statistically different (P = 0.0006).

CONCLUSIONS

In our study, 3 g/day oral GCV did not significantly reduce CMV disease incidence, but there was a suggestion of a death-rate reduction. Furthermore, results suggest that oral GVC decreased risk of CMV disease in patients not prescribed ddI, and increased risk in those prescribed ddI. For the CMV endpoint, our study differs markedly from the only similar study, although for the death endpoint, a combined analysis of studies indicated significant reduction in death rate.

Colorectal epithelial cell proliferative kinetics and risk factors for colon cancer in sporadic adenoma patients

Colorectal epithelial cell proliferative kinetics are altered in patients at increased risk for colon cancer: proliferation rates [labeling index (LI)] are higher and there is a shift of the proliferative zone from one confined to the lower 60% of the colonic crypt to one that includes the entire crypt (higher phi(h)). To assess factors associated with LI and phi(h), we performed a cross-sectional analysis using baseline rectal mucosal biopsies from sporadic adenoma patients participating in a chemoprevention trial. Biopsies (taken without preparatory cleansing) were taken 10 cm above the level of the anus, and proliferation was assessed by detection of endogenous S-phase-associated proliferating cell nuclear antigen by immunohistochemical methods. High-quality, scorable biopsies were obtained for 115 patients, and using analysis of covariance and multiple linear regression, the LI and phi(h) were evaluated in relation to diet and other lifestyle factors, demographics, anthropometrics, family history of colon cancer, and polyp history. Statistically significant findings included the following: (a) The LI for those in the upper versus the lowest tertile of vegetable and fruit consumption was, proportionately, 35% lower (3.4% versus 5.3%; P < 0.001); for vitamin supplement users versus nonusers, it was 36% lower (3.3 versus 5.2%; P < 0.001); for recurrent versus incident polyp patients, it was 36% higher (6.2 versus 4.0%; P < 0.001); and for those with rectal polyps only versus those with colon polyps only, it was 28% higher (6.0 versus 4.3%; P = 0.05); and (b) the phi(h) for those in the upper versus the lowest tertile of sucrose consumption was, proportionately, 48% higher (7.1% versus 3.7%; P = 0.01). These results indicate that (a) colorectal epithelial cell proliferation rates are higher in recurrent adenoma patients than in incident adenoma patients and in patients with rectal adenomas only versus those with colon adenomas only, but they are lower in patients with higher intakes of vegetables and fruit and in those who take vitamin/mineral supplements, and (b) the distribution of proliferating cells is shifted toward more inclusion of the upper 40% of the crypt in patients with higher intakes of sucrose. The pattern of positive, negative, and null associations of potential risk factors with cell proliferation is similar to that commonly found with colonic neoplasms.

Tests of independence for bivariate survival data

We propose two test statistics based on the covariance process of the martingale residuals for testing independence of bivariate survival data. The first test statistic takes the supremum over time of the absolute value of the covariance process, and the second test statistic is a time-weighted summary of the process. We derive asymptotic properties of the two test statistics under the null hypothesis of independence. In addition, we derive the asymptotic distribution of the weighted test and construct optimal weights for contiguous alternatives to independence. Through simulations, we compare the performance of the proposed tests and the inner product of the Savage scores statistics of Clayton and Cuzick (1985, Journal of the Royal Statistical Society, Series A 148, 82-108). These demonstrate that the supremum test is generally more powerful with comparatively little power loss relative to their test when Clayton's family alternative holds, and the weighted test is more powerful when the weight is appropriately chosen.

Hormone replacement therapy is associated with lower risk of adenomatous polyps of the large bowel: the Minnesota Cancer Prevention Research Unit Case-Control Study

Evidence of a role for steroid hormones and reproduction in colon neoplasia remains tantalizing but unclear. Hormone replacement therapy (HRT) has been reported in a number of recent studies to be associated with a reduced risk of colon cancer. A case-control study was undertaken to establish whether HRT is associated with lower risk of adenomatous polyps. This case-control study was undertaken as a project of the Minnesota Cancer Prevention Research Unit. Cases (n = 219) were women, ages 30-74 years with colonoscopy-proven, pathology-confirmed, adenomatous polyps of colon and rectum recruited at Digestive Healthcare PA (Minneapolis, MN). Two control groups were selected: women without polyps at colonoscopy (n = 438) at Digestive Healthcare and age- and zip code-matched women selected from the general community (n = 247). Response rates were 68% among those colonoscoped and 65% among community controls. Parity, age at first live birth, and oral contraceptive use did not distinguish cases from either control group. Multivariate adjusted odds ratios and 95% confidence limits for use of HRT for less than 5 years (compared with never use) among postmenopausal women were 0.52 (0.32-0.85) versus colonoscopy-negative controls and 0.74 (0.44-1.26) versus community controls. For 5 years of use or greater, the corresponding figures were 0.39 (0.23-0.67) and 0.61 (0.34-1.07). These results were not materially different when stratified on body mass index, oophorectomy, hysterectomy, aspirin use, or family history. There is no marked increase in risk even 5 years after cessation of HRT use. HRT appears to lower risk of colorectal adenomatous polyps, suggesting that it acts quite early in the neoplastic process. Mechanisms remain unclear. Reduction of risk of colorectal neoplasia is an additional benefit of postmenopausal HRT.

Consistency in performance among primary care practitioners

OBJECTIVES

The authors studied the consistency of performance of individual physicians to evaluate the identification of outlier practitioners as a strategy for improving patient care.

METHODS

The authors used a data base containing information on 430 practitioners caring for 6,090 patients in 16 group practices. The authors analyzed inter- and intraphysician differences in performance on the basis of review criteria for 8 patient care guidelines. These criteria allowed for a variety of acceptable clinical strategies, incorporated decision tree logic, and included input from participating practitioners. The authors took steps to maximize validity and controlled for potentially confounding characteristics of patients and practitioners. The authors identified outliers, evaluated the significance of differences between outliers and nonoutliers, and studied variations in performance across cases and guidelines in conformance with guidelines.

RESULTS

The authors identified a few statistically significant outliers. Correlations for performance across cases seen by a given physician were low. The highest positive correlation for performance between any pair of guidelines was 0.32.

CONCLUSIONS

The performance of a given practitioner is highly variable from patient to patient and from guideline to guideline. Thus, strategies focusing solely on substandard outliers will miss opportunities to improve performance.

Leadership for quality improvement in group practices

OBJECTIVES

The authors studied influence of physician leaders on their colleagues' performance using data from a randomized, controlled trial of quality assurance interventions in 16 primary care practices.

METHODS

The authors examined whether leaders performed better than their colleagues and looked for evidence of leaders' influence on their colleagues before intervention. As behavioral indicators of each leader's influence for improvement in response to quality assurance interventions, the authors (1) change in the leader's performance score and (2) an index of leader commitment derived from the leader's participation in quality assurance interventions; these indicators were used as covariates in a comparison in experimental practice sites of leaders' and colleagues' mean case performance scores before and after intervention.

RESULTS

Leaders did not outscore their colleagues or influence their colleagues' performance scores before intervention. In response to quality assurance interventions, a leader's change in performance score significantly improved colleagues' score if the leader improved. A positive leader commitment index predicted colleagues' improvement independently of the leader change score.

CONCLUSIONS

As hypothesized, physician leaders, by the example of their behavior, influenced colleagues' performance. However, leaders exerted their influence only after receiving external stimulation for quality improvement.

What makes quality assurance effective? Results from a randomized, controlled trial in 16 primary care group practices

OBJECTIVES

The authors estimate separately contributions of each component intervention to overall effectiveness of quality assurance cycles used to improve practice performance.

METHODS

In a randomized, controlled trial, experimental cycles of quality assurance were conducted for eight patient-care guidelines, with two experimental cycles assigned to each of 16 group practices. For three separate interventions per cycle, practitioners: (1) were notified of the name of the experimental guideline, (2) discussed criteria of conformance to the guideline, and (3) received feedback on performance. Actions taken in response to interventions were documented. Using medical records data for a baseline year and for 3 months after each intervention and an additional 9 months, the authors scored each practice for conformance to two experimental guidelines and to control guidelines.

RESULTS

For all patient-care guidelines combined, and for four of five guidelines showing improvement, knowledge of guidelines and review criteria alone produced no change. After feedback, performance improved and improvement persisted for at least 9 months. The number of corrective actions implemented contributed significantly to effectiveness of quality assurance.

CONCLUSIONS

Feedback to providers of data on their performance is a more powerful stimulus for quality improvement than is knowledge of guidelines or discussion of review criteria.

Healthy worker effect in a longitudinal study of one-second forced expiratory volume (FEV1) and chronic exposure to granite dust

BACKGROUND

Low level effects of granite dust on one-second forced expiratory volume (FEV1) are estimated in 618 Vermont granite workers followed for 5 years with annual pulmonary function tests. Reduced pulmonary function has already been reported for the subset of subjects lost to follow-up (dropouts) suggesting possible bias in analyses based only on survivors.

METHOD

Healthy worker selection bias is directly assessed by comparing the dose-response associations between survivors who remained in the study for the full 5-year observation period and the dropouts.

RESULTS

The 353 survivors had an FEV1 of 96% of predicted at baseline and were losing FEV1 at an average rate of 44 ml/yr. No association was found in this group between the rate of FEV1 decline and lifetime dust exposure. However, the 265 workers with incomplete follow-up, 'dropouts', had a lower FEV1 at baseline (94%) and were losing FEV1 at an average rate of 69 ml/yr. The dose-response parameter in this group was estimated to be 4 ml/yr loss per mg/m3-year and was statistically significant.

CONCLUSIONS

These results provide an illustration of bias due to the healthy worker effect and an example of the failure to detect a true work-related health effect in a study based only on a 'survivor' population.

Inferences on the association parameter in copula models for bivariate survival data

We investigate two-stage parametric and two-stage semi-parametric estimation procedures for the association parameter in copula models for bivariate survival data where censoring in either or both components is allowed. We derive asymptotic properties of the estimators and compare their performance by simulations. Both parametric and semi-parametric estimators of the association parameter are efficient at independence, and the parameter estimates in the margins have high efficiency and are robust to misspecification of dependency structures. In addition, we propose a consistent variance estimator for the semi-parametric estimator of the association parameter. We apply the proposed methods to an AIDS data set for illustration.

Calcium and colorectal epithelial cell proliferation in sporadic adenoma patients: a randomized, double-blinded, placebo-controlled clinical trial

BACKGROUND

The kinetics of colorectal epithelial cell proliferation is altered in patients at increased risk for colon cancer. Calcium administration ameliorates such proliferative changes in rodents. Findings in preliminary clinical trials have suggested similar effects in humans.

PURPOSE

A randomized, double-blind, placebo-controlled, clinical trial was designed to determine whether calcium supplementation will reduce the colorectal epithelial cell proliferation rate and normalize the distribution of proliferating cells within colorectal crypts (i.e., shift the zone of proliferation from the entire crypt to the lower 60% of the crypt, which is thought to be the normal proliferative zone of the crypt) in patients with sporadic adenomas.

METHODS

Sporadic adenoma patients (n = 193) were treated with placebo (n = 66), 1.0 g calcium (n = 64), or 2.0 g calcium (n = 63) daily for 6 months. Rectal mucosa biopsy specimens were obtained at base line and at 1-, 2-, and 6-month follow-up. Cell proliferation was measured by detection of S-phase-associated proliferating cell nuclear antigen by immunohistochemical methods. The cell proliferation rate, called labeling index (LI), was calculated as the proportion of labeled cells in the crypts. The deviation of the proliferative zone from the normal location in the lower 60% of the crypt was calculated as the proportion of labeled cells in the upper 40% of the crypt, called distributional index (phi h). The effects of calcium treatment on the LI and phi h were expressed as relative effects--(calcium follow-up/calcium base line)/(placebo follow-up/placebo base line). Calculations and inference testing of the relative effects were accomplished using a repeated-measures mixed model on log-transformed LI and phi h values. All statistical tests were two-sided.

RESULTS

Scorable biopsy specimens were obtained on 170 patients at base line, 164 at 1 month, 161 at 2 months, and 163 at 6 months. The difference in the change in the LI between the combined calcium groups and the placebo group was insignificant, with a relative effect of calcium versus placebo of 0.97 (P = .87). However, for the phi h, the relative effect of calcium versus placebo was 0.50 (P = .05) in the combined calcium groups, 0.56 (P = .16) in the 1.0-g calcium group, and 0.44 (P = .05) in the 2.0-g calcium group.

CONCLUSIONS

Calcium supplementation normalizes the distribution of proliferating cells without affecting the proliferation rate in the colorectal mucosa of sporadic adenoma patients.

IMPLICATIONS

These results support further study of whether alterations in colon cell proliferative kinetics represent true intermediate steps in colon carcinogenesis that can be used to investigate the etiology and prevention of, and whether a higher calcium consumption can reduce the risk of, colon cancer.

Long-term effects of reduced renal mass in humans

The long-term risks of kidney donation have not been well defined. We carried out meta-analysis of investigations that examined the long-term effects of reduced renal mass in humans. We used multiple linear regression to combine studies and adjust for differences in the duration of follow-up, the reason for reduced renal mass, the type of controls, age and gender. We analyzed 48 studies with 3124 patients and 1703 controls. Unilateral nephrectomy caused a decrement in glomerular filtration rate (-17.1 ml/min; 95% confidence interval -20.2 to -14.0 ml/min) that tended to improve with each 10 years of follow-up (1.4 ml/min/decade; 0.3 to 2.4 ml/min/decade). Patients with single kidneys had small, progressive increases in proteinuria (76 mg/day/decade; 52 to 101 mg/day/decade), but proteinuria was negligible after nephrectomy for trauma or kidney donation. Nephrectomy did not affect the prevalence of hypertension, but there was a small increase in systolic blood pressure (2.4 mm Hg; -0.3 to 5.1 mm Hg, P > 0.05) which rose further with duration of follow-up (1.1 mm Hg/decade; 0.0 to 2.2 mm Hg/decade). Diastolic blood pressure was higher after nephrectomy (3.1 mm Hg; 1.8 to 4.4 mm Hg), but this increment did not change with duration of follow-up. Thus, in normal individuals, unilateral nephrectomy does not cause progressive renal dysfunction, but may be associated with a small increase in blood pressure.

Lipid-lowering therapy in patients with renal disease

A growing number of clinical trials have examined the effects of different lipid lowering strategies in patients with renal disease. We carried out a meta-analysis to compare and contrast the relative efficacy of various antilipemic therapies in different renal disease settings. Studies that investigated one or more therapies designed to lower serum lipids were combined using weighted multiple linear regression. The analysis adjusted treatment effects for differences in baseline lipid levels and possible placebo effects. The results showed that antilipemic therapies generally had similar effects on lipids in different renal disease settings. In nephrotic syndrome the greatest and most consistent reductions in low density lipoprotein cholesterol (LDL) were seen with 3-hydroxy-3-methylglutaryl co-enzyme A (HMG-CoA) reductase inhibitors (regression coefficient with 95% confidence interval in mg/dl = -63, -79 to -46). Similar results were seen for LDL in renal transplant (-51, -57 to -45), renal insufficiency (-62, -82 to -42), hemodialysis (-65, -80 to -50) and continuous ambulatory peritoneal dialysis (CAPD) patients (-84, -104 to -64). Fibric acid analogues had less effect on LDL, but caused greater reductions in triglycerides: -132, -178 to -87, in nephrotic syndrome; -69, -93 to -45 in transplant: -107, -169 to -45 in renal insufficiency; -72, -120 to -24 in hemodialysis; and -96, -162 to -30 in CAPD. In general, the effects of diet and other therapies were less consistent. Despite possible limitations of this meta-analysis, the results provide a useful framework for choosing antilipemic therapy, and point to areas for future long-term studies examining the safety and efficacy of lipid lowering strategies in patients with renal disease.

Long-term effects of antihypertensive agents on proteinuria and renal function

BACKGROUND

Although many studies have examined the effects of antihypertensive agents on proteinuria and glomerular filtration rate in patients with kidney disease, many questions remain unresolved. These questions include whether the effects of agents differ, whether their effects are similar in diabetic and nondiabetic patients with renal disease, and whether the effects of any agents are independent of blood pressure reductions.

METHODS

We conducted a meta-analysis of studies obtained with MEDLINE and bibliographies from comprehensive reviews but included only investigations with follow-up times of at least 6 months. We combined data (1) in an analysis of randomized controlled trials, (2) in a separate univariate analysis of controlled and uncontrolled trials, and (3) using weighted multiple linear regression.

RESULTS

In 14 randomized controlled trials, angiotensin-converting enzyme inhibitors caused a greater decrease in proteinuria (pooled mean [95% confidence intervals], -0.51[-0.68 to -0.35], ln [treatment/control]), improvement in glomerular filtration rate (0.13 mL/min per month [0.10 to 0.16 mL/min per month]), and decline in mean arterial pressure (-4.0 mm Hg [-4.9 to -3.0 mm Hg]) compared with controls. In a multivariate analysis of controlled and uncontrolled trials, each 10-mm Hg reduction in blood pressure decreased proteinuria (regression coefficient [95% confidence interval] -0.14 [-0.22 to -0.06] ln [after/before]), but angiotensin-converting enzyme inhibitors (-0.45 [-0.58 to -0.32]) and nondihydropyridine calcium antagonists (-0.38 [-0.70 to -0.06]) were associated with additional declines in proteinuria that were independent of blood pressure changes and diabetes. Each 10-mm Hg reduction in blood pressure caused a relative improvement in glomerular filtration rate (0.18 mL/min per month [0.04 to 0.31 mL/min per month]), but among diabetic patients there was a tendency for dihydropyridine calcium antagonists to cause a relative reduction in glomerular filtration rate (-0.68 mL/min per month [-1.31 to -0.04 mL/min per month]).

CONCLUSIONS

Long-term beneficial effects of antihypertensive agents on proteinuria and glomerular filtration rate are proportional to blood pressure reductions and are similar in diabetic and nondiabetic patients with renal disease. In addition, angiotensin-converting enzyme inhibitors, and possibly nondihydropyridine calcium antagonists, have additional beneficial effects on proteinuria that are independent of blood pressure reductions.

Effects of antihypertensive therapy on serum lipids

PURPOSE

To compare and contrast the effects of antihypertensive agents on serum lipids and blood pressure in different patient populations.

DATA SOURCES

A MEDLINE search and bibliographies from recent comprehensive reviews were used to identify trials that provided sufficient data to calculate the change in one or more serum lipid values measured before and after antihypertensive therapy.

STUDY SELECTION

474 controlled and uncontrolled clinical trials investigated the effects of 85 antihypertensive agents on lipids and blood pressure in more than 65,000 patients.

DATA EXTRACTION

Data on triglyceride and total, low-density lipoprotein (LDL), and high-density lipoprotein (HDL) cholesterol levels; blood pressure; patient characteristics; and study design.

DATA SYNTHESIS

Differences in the effects of agents, adjusted for differences in patient populations and study design, were examined using multiple linear regression analysis that was weighted by study quality and inverse variance. Diuretics caused relative increases in cholesterol levels (regression coefficient = 0.13 mmol/L; 95% CI, 0.09 to 0.18 mmol/L) that were greater with higher doses (additional effect of high dose, 0.12 mmol/L; CI, 0.04 to 0.20 mmol/L) and were worse in blacks than in nonblacks (additional effect in blacks, 0.13 mmol/L; CI, 0.01 to 0.26 mmol/L). Beta-blockers caused increases in triglyceride levels (0.35 mmol/L; CI, 0.31 to 0.39 mmol/L) that were substantially smaller for agents with intrinsic sympathomimetic activity (amelioration of beta-blocker increase, -0.21 mmol/L; CI, -0.27 to -0.16 mmol/L). When combined with cardioselectivity, beta-blockers with intrinsic sympathomimetic activity favorably affected lipids and reduced both total (-0.14 mmol/L; CI, -0.24 to -0.04 mmol/L) and LDL cholesterol levels (-0.17 mmol/L; CI, -0.28 to -0.07 mmol/L). alpha-Blockers beneficially affected total cholesterol (-0.23 mmol/L; CI, -0.28 to -0.18 mmol/L), LDL cholesterol (-0.20 mmol/L; CI, -0.25 to 0.15 mmol/L), triglycerides (-0.07 mmol/L; CI, -0.11 to -0.03 mmol/L), and, in younger persons, HDL cholesterol (0.02 mmol/L; 0.01 to 0.04 mmol/L). Converting enzyme inhibitors reduced triglycerides (-0.07 mmol/L; CI, -0.12 to -0.02 mmol/L), and, in patients with diabetes, total cholesterol (-0.22 mmol/L; CI, -0.34 to -0.10 mmol/L). Vasodilators reduced total (-0.22 mmol/l; CI, -0.30 to -0.10 mmol/L) and LDL cholesterol (-0.22 mmol/L; CI, -0.29 to -0.11 mmol/L) and increased HDL cholesterol (0.06 mmol/L; CI, 0.02 to 0.09 mmol/L).

CONCLUSION

With the exception of calcium antagonists, nearly all antihypertensive agents affect serum lipids. These effects differ among patient populations.

Assessing gamma frailty models for clustered failure time data

Proportional hazards frailty models use a random effect, so called frailty, to construct association for clustered failure time data. It is customary to assume that the random frailty follows a gamma distribution. In this paper, we propose a graphical method for assessing adequacy of the proportional hazards frailty models. In particular, we focus on the assessment of the gamma distribution assumption for the frailties. We calculate the average of the posterior expected frailties at several followup time points and compare it at these time points to 1, the known mean frailty. Large discrepancies indicate lack of fit. To aid in assessing the goodness of fit, we derive and estimate the standard error of the mean of the posterior expected frailties at each time point examined. We give an example to illustrate the proposed methodology and perform sensitivity analysis by simulations.

Survival and disease progression according to gender of patients with HIV infection. The Terry Beirn Community Programs for Clinical Research on AIDS

OBJECTIVE

To compare disease progression and mortality between women and men infected with human immunodeficiency virus (HIV).

DESIGN

Multicenter cohort.

SETTING

Seventeen community-based centers participating in the Terry Beirn Community Programs for Clinical Research on AIDS (CPCRA).

PATIENTS

A total of 768 women and 3779 men enrolled in one or more of 11 protocols between September 7, 1990, and September 30, 1993.

MAIN OUTCOME MEASURES

Survival and opportunistic events.

RESULTS

The median CD4+ cell count at enrollment into the cohort was 0.240 x 10(9)/L (240/microL) for women and 0.137 x 10(9)/L for men (P < .001). Compared with men, women were younger (36 vs 38 years), more likely to be African American or Hispanic (78% vs 44%), and more likely to have reported a history of injection drug use (49% vs 27%). Women had been followed up for a median of 14.5 months and men for 15.5 months. The adjusted relative risk (RR) for death among women compared with men was 1.33 (95% confidence interval [CI], 1.06 to 1.67; P = .01) and for disease progression (including death) was 0.97 (95% CI, 0.82 to 1.15; P = .72). Women were at increased risk for bacterial pneumonia (RR, 1.38; 95% CI, 1.05 to 1.92) and at reduced risk for the development of Kaposi's sarcoma (RR, 0.16; 95% CI, 0.04 to 0.65) and oral hairy leukoplakia (RR, 0.54; 95% CI, 0.31 to 0.94). The increased risk of death and bacterial pneumonia for women compared with men was primarily evident among those with a history of injection drug use (RR, 1.68 for death, 95% CI, 1.20 to 2.35, P = .003; RR, 1.53 for bacterial pneumonia, 95% CI, 1.03 to 2.29, P = .04). Among patients without a history of disease progression at entry, death was the first event reported for more women than men (27.5% vs 12.2%).

CONCLUSIONS

Compared with men, HIV-infected women in the CPCRA were at increased risk of death but not disease progression. Risks of most incident opportunistic diseases were similar for women and men; however, women were at an increased risk of bacterial pneumonia. These findings may reflect differential access to health care and standard treatments or different socioeconomic status and social support for women compared with men.

Empirical Bayes methods for stabilizing incidence rates before mapping

The epidemiologic utility of mapping and ranking incidence rates is often questioned owing to instability of the observed incidence values in areas with small populations. Spurious fluctuations in the observed rates caused by this instability can mask true spatial and temporal trends in risk. To produce maps with the required level of geographic resolution, yet based on reliable estimates, it is desirable to reduce the random variation in the observed rates before mapping. In this paper, we describe the empirical Bayes approach for obtaining stabilized incidence estimates. We begin by deriving Bayes rate estimators and then illustrate how using the observed rates to estimate unknown distributional information leads to the empirical Bayes formulation. A drawback of the approach is that the histogram of the empirical Bayes rate estimates may be narrower than the true distribution of risk. We outline a constrained empirical Bayes approach that produces improved estimators for the true distribution of the unknown rates. We include discussions of relevant previous applications of empirical Bayes methods to rate mapping problems and an evaluation of the strengths and weaknesses of the approach.

Statistical analysis of proliferative index data in clinical trials

Measurements of proliferative activity in colonic epithelial cells are being used as surrogate endpoints in clinical trials for colon cancer prevention. Proliferative index data exemplify an important type of clinical trial endpoint. The outcome variable is a proportion in which the denominator is an ancillary statistic and in which measurement error and technician judgement are important sources of variability. The paper proposes a statistical model for a repeated measures clinical trial with this type of endpoint, in the context of proliferative activity data. The model is a two-stage random effects linear model in the log scale. In addition to fixed effects covariates, it explicitly incorporates two major sources of variability: the number of epithelial cells counted and the reader effect. Although the resulting likelihood is complicated, one can fit an approximate likelihood with minimal loss of efficiency using standard packages. We apply the model to a pilot randomized clinical trial.

A constrained empirical Bayes estimator for incidence rates in areas with small populations

Maps that show the geographic distribution of incidence rates can be useful tools for analysing spatial variation in mortality and morbidity. To attain the necessary geographic resolution, however, production of such maps often requires estimation of incidence in areas with small populations where the observed rates may be highly unstable. Manton et al. have presented an empirical Bayes stabilization procedure in which the observed rate is combined with an area-specific estimate of the underlying incidence. The approach allows for the mapping of outcomes with varied and possibly unknown etiologies without necessitating covariate dependent modelling of the expected rate. The empirical distribution of a collection of these estimates, however, may not provide an adequate description of the dispersion among the true rates. As a result, decisions based on the histogram of the empirical Bayes estimates may be suspect. We propose a modified version of the approach in which the mean and sample variance of the ensemble of estimates are constrained to equal the appropriate moments of the posterior distribution. The resulting collection of constrained empirical Bayes estimators has nearly the stability of the unconstrained approach and provides an improved estimator of the true rate distribution. We illustrate use of the estimator by producing stabilized county-level maps of U.S. fire- and burn-related mortality rates and validate the analytic results using a simulation analysis.

Association between survival time and ordinal covariates

An application of the method of rank correlation is proposed for testing independence between a censored survival time and an ordinal covariate. The test statistic counts the number of concordances minus the number of discordances at each time with event(s) and adds across times; it is expressible as a score statistic within the proportional hazards framework. The proposed test includes, as a special case, a generalization of Jonckheere's test against ordered alternatives and as applied to the analysis of categorical data, it can be seen as a generalization of the Mantel-Haenszel procedure.

Chronic conditions that lead to functional limitation in the elderly

BACKGROUND

The role of chronic medical conditions in elderly persons' loss of functional abilities is intuitively important but not well defined. This analysis was designed to identify chronic medical conditions that lead to the development of severe functional limitation.

METHODS

Functionally intact members of a multistage probability sample (n = 6,862) of all noninstitutionalized U.S. civilians age 70 years or older were interviewed in 1984. Based on data from the National Death Index and from follow-up telephone interviews in 1988 with survivors, subjects were classified as functionally intact, functionally limited (unable to perform one or more of seven essential activities), or decreased.

RESULTS

After adjusting for the effects of exercise habits and demographic, socioeconomic, and psychosocial factors, we found that the best predictors of the development of functional limitation were cerebrovascular disease (OR = 2.14; 95% CL = 1.16, 3.98) and arthritis (OR = 1.51; 95% CL = 1.08, 2.11). The contribution of coronary artery disease also approached statistical significance (OR = 1.49; 95% CL = 0.99, 2.27).

CONCLUSION

In the future, the primary prevention or effective treatment of cerebrovascular disease, arthritis, and possibly coronary artery disease may produce a modest reduction in the incidence of severe functional limitation.

Calcium and colorectal epithelial cell proliferation: a preliminary randomized, double-blinded, placebo-controlled clinical trial

BACKGROUND

Colonic epithelial cell proliferation is increased in patients at high risk for colon cancer. Calcium administration has ameliorated the proliferative changes in rodents, and findings in small, uncontrolled clinical trials have suggested similar effects in humans.

PURPOSE

This preliminary, double-blind, randomized clinical trial was designed 1) to investigate whether supplemental calcium will reduce colonic epithelial cell proliferation in patients with sporadic adenomas who consume a high-fat, Western-style diet; 2) to determine the sample size (number of scorable crypts per person) needed to achieve adequate statistical power; and 3) to evaluate the feasibility of full-scale clinical trials.

METHODS

Twenty-one sporadic adenoma patients were treated daily with placebo or 1200 mg of supplemental calcium. To determine colonic epithelial cell proliferation, we used tritiated thymidine labeling of colon crypt epithelial cells in rectal biopsy specimens and calculated the percentage of labeled cells (labeling index [LI]). Two pathology technician "readers" independently scored each specimen, and inter-reader reliability was determined. Subjects remained on their usual diet during the study, and intake of calories, calcium, total fat, and vitamin D did not differ substantially among them. We calculated curves for statistical power to determine the number of scorable crypts needed per person for detection of a statistically significant difference (P < .05) of 1.0% in mean LI.

RESULTS

The pooled baseline LI was 4.7%. In the calcium-treated group, the LI increased 0.6% (proportional increase, 12.8%); in the placebo-treated group, it decreased 0.5% (proportional decrease, 10.6%). The difference between change in the mean LI from baseline to 8 weeks' follow-up in the placebo group versus the calcium group was not statistically significant. The intraclass correlation coefficient for inter-reader reliability for the baseline LI was .66. Analyses indicated scoring eight crypts sufficient for estimates of the LI adequate for between-group comparisons, a level achieved in 81% of biopsy specimens.

CONCLUSIONS

Calcium carbonate supplements delivering 1200 mg elemental calcium daily may not decrease colonic epithelial cell proliferation over an 8-week period in sporadic adenoma patients. In future trials measuring the LI, consideration should be given to ensuring adequate numbers of scorable crypts and to the impact of inadequate biopsy procedures, labeling failure, reader reliability, and participant withdrawal. Our findings support the feasibility of a full-scale clinical trial to further study the relationships among dietary calcium, colonic epithelial cell proliferation, and colorectal cancer.

Random effects models with non-parametric priors

We discuss the performance of non-parametric maximum likelihood (NPML) estimators for the distribution of a univariate random effect in the analysis of longitudinal data. For continuous data, we analyse generated and real data sets, and compare the NPML method to those that assume a Gaussian random effects distribution and to ordinary least squares. For binary outcomes we use generated data to study the moderate and large-sample performance of the NPML compared with a method based on a Gaussian random effect distribution in logistic regression. We find that estimated fixed effects are compatible for all approaches, but that appropriate standard errors for the NPML require adjusting the likelihood-based standard errors. We conclude that the non-parametric approach provides an attractive alternative to Gaussian-based methods, though additional evaluations are necessary before it can be recommended for general use.

Using empirical Bayes methods in biopharmaceutical research

A compound sampling model, where a unit-specific parameter is sampled from a prior distribution and then observed are generated by a sampling distribution depending on the parameter, underlies a wide variety of biopharmaceutical data. For example, in a multi-centre clinical trial the true treatment effect varies from centre to centre. Observed treatment effects deviate from these true effects through sampling variation. Knowledge of the prior distribution allows use of Bayesian analysis to compute the posterior distribution of clinic-specific treatment effects (frequently summarized by the posterior mean and variance). More commonly, with the prior not completely specified, observed data can be used to estimate the prior and use it to produce the posterior distribution: an empirical Bayes (or variance component) analysis. In the empirical Bayes model the estimated prior mean gives the typical treatment effect and the estimated prior standard deviation indicates the heterogeneity of treatment effects. In both the Bayes and empirical Bayes approaches, estimated clinic effects are shrunken towards a common value from estimates based on single clinics. This shrinkage produces more efficient estimates. In addition, the compound model helps structure approaches to ranking and selection, provides adjustments for multiplicity, allows estimation of the histogram of clinic-specific effects, and structures incorporation of external information. This paper outlines the empirical Bayes approach. Coverage will include development and comparison of approaches based on parametric priors (for example, a Gaussian prior with unknown mean and variance) and non-parametric priors, discussion of the importance of accounting for uncertainty in the estimated prior, comparison of the output and interpretation of fixed and random effects approaches to estimating population values, estimating histograms, and identification of key considerations in the use and interpretation of empirical Bayes methods.

Variance components and their implications for statistical information in medical data

We discuss general issues concerning the design and analysis of medical experiments involving repeated measures or hierarchical groupings of subjects within larger study units. Depending on the types of questions being investigated, the correlations induced by clustering can have dramatic impact on the effective sample size. The unique aspects of such experiments must be accounted for during analysis and during interpretation of the results. We illustrate these issues by using a variance components model to investigate the role of leadership in medical practice.

Forecasting the number of future disabled elderly using Markovian and mathematical models

The accuracy of forecasting the number of future disabled elderly people depends on the accuracy of projecting mortality rates and the rates of transition to and from functional disability. We describe a new two-step method for constructing mathematical models that project these future rates dynamically. (1) A Markovian model of elders' transitions between functional states is specified. (2) A mathematical model of the probability of each transition is created. We conducted pilot studies of the fundamental mathematical processes of this method using data from the Longitudinal Study of Aging. First we constructed prototypic mathematical models of the probabilities of remaining functionally able and of making transitions to disability and to death within 2 years. Then we used these models to project hypothetical rates of transition for white women of selected ages, morbidity ratings and health statuses.

Assessing, accommodating, and interpreting the influences of heterogeneity

Heterogeneity, ranging from measurement error to variation among individuals or regions, influences all levels of data collected for risk assessment. In its role as a nemesis, heterogeneity can reduce the precision of estimates, change the shape of a population model, or reduce the generalizability of study results. In many contexts, however, heterogeneity is the primary object of inference. Indeed, some degree of heterogeneity in excess of a baseline amount associated with a statistical model is necessary in order to identify important determinants of response. This report outlines the causes and influences of heterogeneity, develops statistical methods used to estimate and account for it, discusses interpretations of heterogeneity, and shows how it should influence study design. Examples from dose-response modeling, identification of sensitive individuals, assessment of small area variations and meta analysis provide applied contexts.

Longitudinal and cross-sectional estimates of pulmonary function decline in never-smoking adults

This paper describes methods for simultaneous cross-sectional and longitudinal analysis of repeated measurements obtained in cohort studies with regular examination schedules, then uses these methods to describe age-related changes in pulmonary function level among nonsmoking participants in the Six Cities Study, a longitudinal study of air pollution and respiratory health conducted between 1974 and 1983 in Watertown, Massachusetts; Kingston and Harriman, Tennessee; St. Louis, Missouri; Steubenville, Ohio; Portage, Wisconsin; and Topeka, Kansas. The subjects, initially aged 25-74, were examined on three occasions at 3-year intervals. Individual rates of loss increased more rapidly with age than predicted from the cross-sectional model. For example, for a male of height 1.75 m, the cross-sectional model predicted an increase in the annual rate of loss of FEV1 from 23.7 ml/yr at age 25 to 39.0 ml/yr at age 75, while the longitudinal model gave rates of loss increasing from 12.9 ml/yr at age 25 to 58.2 ml/yr at age 75. These results contrast with those of other studies comparing longitudinal and cross-sectional estimates of pulmonary function loss.

Performance quality, gender, and professional role. A study of physicians and nonphysicians in 16 ambulatory care practices

The quality of medical care has rarely been evaluated in relation to practitioner or patient gender. Moreover, comparisons between physicians and nonphysicians typically are confounded by practitioner gender. In this study gender and professional role effects were analyzed separately for 162 male and female staff physicians, 191 male and female residents, and 73 female nonphysicians delivering adult and pediatric primary care in 16 ambulatory care practices. Analyses addressed influences of patient and practitioner gender as well as differences between physicians and nonphysicians. Results showed that female staff physicians performed better than male staff physicians for cancer screening in women by breast examination and Pap smears, but that female residents performed worse than male residents for urinary tract infections in children. Patient gender effects occurred for two tasks; for these, superior care was rendered to the gender with higher prevalence for the condition (girls for urinary tract infections, boys for otitis media). The results are considered in the context of the gender-relevance of particular medical tasks or conditions. Comparisons between physicians and nonphysicians were limited to female practitioners. Comparable or superior performance for nonphysicians was found for all tasks but one (cancer screening in women).

Diet, plasma levels of beta-carotene and alpha-tocopherol, and risk of malignant melanoma

Dietary intake and the plasma levels of retinol, alpha-tocopherol, lycopene, alpha-carotene, and beta-carotene for 204 cases with malignant melanoma were compared with those of 248 controls. Cases and controls were patients 18 years of age or older making their first visit to a dermatology subspecialty clinic for pigmented lesions from July 1, 1982 to September 1, 1985. Intakes of nutrients were estimated using a semiquantitative food frequency questionnaire. No significant associations with malignant melanoma were observed for higher plasma levels of lycopene, retinol, or alpha-carotene in logistic regression analyses after controlling for age, sex, plasma lipids, and known constitutional risk factors (hair color and ability to tan). In similar models, the odds ratio comparing the highest with the lowest quintile was 0.9 (95% confidence interval (CI) 0.5-1.5) for plasma beta-carotene, 0.7 (95% CI 0.5-1.3) for plasma alpha-tocopherol, 0.7 (95% CI 0.4-1.2) for carotene intake, and 0.7 (95% CI 0.4-1.3) for total vitamin E intake. A trend toward reduced risk of melanoma was observed for increasing intake of iron (not including supplements); this was related to the more frequent consumption of baked goods, such as cake, among controls. Alcohol consumption was positively associated with risk of melanoma (chi for trend = 2.1, p = 0.03); the odds ratio for consumption of over 10 g/day compared with persons with no alcohol intake was 1.8 (95% CI 1.0-3.3).