Richard P. Waterman

Fixed–Effects Negative Binomial Regression Models

This paper demonstrates that the conditional negative binomial model for panel data, proposed by Hausman, Hall, and Griliches (1984), is not a true fixed-effects method. This method—which has been implemented in both Stata and LIMDEP—does not in fact control for all stable covariates. Three alternative methods are explored. A negative multinomial model yields the same estimator as the conditional Poisson estimator and hence does not provide any additional leverage for dealing with over-dispersion. On the other hand, a simulation study yields good results from applying an unconditional negative binomial regression estimator with dummy variables to represent the fixed effects. There is no evidence for any incidental parameters bias in the coefficients, and downward bias in the standard error estimates can be easily and effectively corrected using the deviance statistic. Finally, an approximate conditional method is found to perform at about the same level as the unconditional estimator.

Patient education model for increasing handwashing compliance.

BACKGROUND A review of the literature on handwashing has documented the absence of research on the education of the patient as an intervention model for changing staff behavior regarding handwashing compliance. The primary objective of this project was to conduct a prospective control study of the effect of patient handwashing education on staff compliance with handwashing. METHOD A prospective, controlled, 6-week intervention/control study was performed in 4 community hospitals in South Jersey. Each hospital served as its own control. Patients were educated within 24 hours of admission about the importance of asking their health care workers to wash their hands. Soap usage and handwashing was calculated by bed-days. Patient follow-up was conducted through telephone interviews 2 weeks after discharge. RESULTS The patient handwashing education model increased soap usage by health care workers an average of 34% (P =.021); this increase was consistent across hospitals regardless of the initial soap usage rates. Of the patients interviewed, 81% read the materials provided, 57% asked health care workers whether they had washed their hands, and 81% of this 57% said they received positive responses. CONCLUSIONS For the first time, our findings document that education of patients regarding their role in monitoring handwashing compliance among health care workers can increase soap usage and handwashing and provide sustainable reinforcement of handwashing principles for health care workers.

Validation of venous leg ulcer guidelines in the United States and United Kingdom

BACKGROUND Venous leg ulcers account for 85% of all lower-extremity ulcers, with treatment costs of 3 billion dollars and loss of 2 million workdays per year. The purpose of this study was to validate the clinical efficacy and cost effectiveness of multidisciplinary guidelines for the diagnosis and treatment of venous leg ulcers. METHODS Eighty (40 retrospective, 40 prospective) patients from the United States and United Kingdom were enrolled. RESULTS United States patients were 6.5 times and United Kingdom 2 times more likely to heal if a guideline was followed (P <0.001). A significant decrease was noted in healing time for both the United States and United Kingdom (P <0.01), and the median cost decreased significantly when the guideline was followed (P <0.01). CONCLUSIONS Implementation of a guideline for diagnosis and treatment of venous leg ulcers resulted in improvement in diagnosis, decrease in healing time, and an increase in healing rates resulting in lower costs.

Hand Hygiene Compliance Rates in the United States—A One-Year Multicenter Collaboration Using Product/Volume Usage Measurement and Feedback

Hand hygiene (HH) is the single most important factor in the prevention of health care-acquired infections. The 3 most frequently reported methods of measuring HH compliance are: (1) direct observation, (2) self-reporting by health care workers (HCWs), and (3) indirect calculation based on HH product usage. This article presents the results of a 12-month multicenter collaboration assessing HH compliance rates at US health care facilities by measuring product usage and providing feedback about HH compliance. Our results show that HH compliance at baseline was 26% for intensive care units (ICUs) and 36% for non-ICUs. After 12 months of measuring product usage and providing feedback, compliance increased to 37% for ICUs and 51% for non-ICUs. (ICU, P = .0119; non-ICU, P < .001). HH compliance in the United States can increase when monitoring is combined with feedback. However, HH still occurs at or below 50% compli- ance for both ICUs and non-ICUs.

Consumer Attitudes About Health Care-Acquired Infections and Hand Hygiene

Mandatory reporting and disclosure of health care-acquired infections have resulted in controversy over the perceived notion that consumers will not understand how to interpret data and that such information may negatively influence utilization of hospitals. The objective was to determine consumers’ attitudes about health care-acquired infections, hand hygiene practices, and patient empowerment. A telephone survey based on a random digit dialing sample of all households in the United States was conducted. Consumers were asked about choosing a hospital, hand hygiene practices, and health care-acquired infections. Some 94% of respondents rated environmental cleanliness as very important. Hospital infection rates would influence decision making for 93% of consumers. Four in 5 consumers said they would ask their health care worker to wash and sanitize his or her hands. Our findings strongly suggest that (1) consumers will use infection data in selecting and/or leaving a hospital system and (2) consumers are ready to be empowered with information to ensure a positive outcome.

Efficiency of projected score methods in rectangular array asymptotics

The paper considers a rectangular array asymptotic embedding for multistratum data sets, in which both the number of strata and the number of within-stratum replications increase, and at the same rate. It is shown that under this embedding the maximum likelihood estimator is consistent but not efficient owing to a non-zero mean in its asymptotic normal distribution. By using a projection operator on the score function, an adjusted maximum likelihood estimator can be obtained that is asymptotically unbiased and has a variance that attains the Cramer-Rao lower bound. The adjusted maximum likelihood estimator can be viewed as an approximation to the conditional maximum likelihood estimator. Copyright 2003 Royal Statistical Society.

Business Analysis Using Regression

Class 1. Fitting Equations to Data.- Efficiency of Cleaning Crews.- Liquor Sales and Display Space.- Managing Benefits Costs.- Predicting Cellular Phone use.- Class 2. Assumptions in Regression Modeling.- The Ideal Regression Model.- Predicting Cellular Phone use, Revisited.- Efficiency of Cleaning Crews, Revisited.- Housing Prices and Crime Rates.- Direct Mail Advertising and Sales.- Housing Construction.- Class 3. Prediction and Confidence Intervals in Regression.- Housing Construction, Revisited.- Liquor Sales and Display Space, Revisited.- Class 4. Multiple Regression.- Automobile Design.- Class 5. Collinearity.- Stock Prices and Market Indices.- Improving Parcel Handling.- Class 6. Modeling Categorical Factors with two Levels.- Employee Performance Study.- Class 7. Modeling Categorical Factors with two or More Levels.- Wage Discrimination, Revisited.- Timing Production Runs.- Class 8. Summary Regression Case.- Executive Compensation.- Using Stepwise Regression for Prediction.- Class 9. Comparing Many Mean Values.- Selecting the Best Vendor.- Headache Pain Relief.- Analysis of Variance and Tests for Linearity.- Class 10. Analysis of Variance with Two Factors.- Package Design Experiment.- Evaluating Employee time Schedules.- Class 11. Modeling a Categorical Response.- The Challenger Disaster.- Marketing Orange Juice.- Class 12. Modeling Time Series.- Predicting Cellular Phone use, Revisited.- Trends in Computer Sales.- Assignments.- Appendix: Use with Minitab.

A Cross-National Study of Adolescent Peer Concordance on Issues of the Future

Adolescents in 4 nations were surveyed about theirperceptions regarding adulthood and perceptions they assumed peers to hold. Central tenets of Dynamic Functional Interaction were tested. This theory of adolescent psychological development argues that adolescents try to see themselves in concordance with age-mates to reduce uncertainties associated with transition to adulthood. Approximately 3,500 individuals participated in the cross-national study that explored the association of nationality, age, and gender with level of perceived concordance. Findings disclosed presence of universals as well as national effects. Significant levels of concordance in each nation, at all ages, and in both genders indicated commonalities in the developmental process irrespective of nationality. Scottish youths perceived themselves in concordance with peers almost 50% more often than did U.S. youths. The concordance pattern with respect to age was different markedly in the Philippines. Overall, age and nation had a strongerassociation with concordance than did gender.

Basic Business Statistics

Overview and Foundations.- Statistical Summaries of Data.- Sources of Variation.- Standard Error.- Confidence Intervals.- Sampling.- Making Decisions.- Designing Tests for Better Comparisons.- Confounding Effects in Tests: A Case Study.- Covariance, Correlation, and Portfolios.- A Preview of Regression.

Simulation Modeling for Cost Estimation

What do a group of statisticians have to contribute to the subject of good costing practices, and in particular the United States Postal Service (USPS) costing methodology? Credible cost estimates require the collection of high quality information as component inputs. Deciding what information to collect lies in the province of the economist, but exactly how to collect that information, when to collect it, how much of it to collect, as well as a significant part of the overall evaluation of the quality of the information itself, lies in the province of the statistician.2