Brian J. Carnahan

National observational study of prescription dispensing accuracy and safety in 50 pharmacies.

OBJECTIVES To measure dispensing accuracy rates in 50 pharmacies located in 6 cities across the United States and describe the nature and frequency of the errors detected. DESIGN Cross-sectional descriptive study. SETTINGS Chain, independent, and health-system pharmacies (located in hospitals or managed care organizations). PARTICIPANTS Pharmacy staff at randomly selected pharmacies in each city who accepted an invitation to participate. INTERVENTION Observation by a pharmacist in each pharmacy for 1 day, with a goal of inspecting 100 prescriptions for dispensing errors (defined as any deviation from the prescribers order). MAIN OUTCOME MEASURE Dispensing errors on new and refill prescriptions. RESULTS Data were collected between July 2000 and April 2001. The overall dispensing accuracy rate was 98.3% (77 errors among 4,481 prescriptions; range, 87.2%-100.0%; 95.0% confidence interval, ± 0.4%). Accuracy rates did not differ significantly by pharmacy type or city. Of the 77 identified errors, 5 (6.5%) were judged to be clinically important. CONCLUSION Dispensing errors are a problem on a national level, at a rate of about 4 errors per day in a pharmacy filling 250 prescriptions daily. An estimated 51.5 million errors occur during the filling of 3 billion prescriptions each year.

Designing safe job rotation schedules using optimization and heuristic search

Job rotation is one method that is sometimes used to reduce exposure to strenuous materials handling; however, developing effective rotation schedules can be complex in even moderate sized facilities. The purpose of this research is to develop methods of incorporating safety criteria into scheduling algorithms to produce job rotation schedules that reduce the potential for injury. Integer programming and a genetic algorithm were used to construct job rotation schedules. Schedules were comprised of lifting tasks whose potential for causing injury was assessed with the Job Severity Index. Each method was used to design four job rotation schedules that met specified safety criteria in a working environment where the object weight, horizontal distance and repetition rate varied over time. Each rotation was assigned to a specific gender/lifting capacity group. Five versions of the integer programming search method were applied to this problem. Each version generated one job rotation schedule. The genetic algorithm model was able to create a population of 437 feasible solutions to the rotation problem. Utilizing cluster analysis, a rule set was derived from the genetic algorithm generated solutions. These rules provided guidelines for designing safe job rotation schedules without the use of a computer. The advantages and limitations of these approaches in developing administrative controls for the prevention of back injury are discussed.

Incorporating Physical Demand Criteria into Assembly Line Balancing

Many assembly line balancing algorithms consider only task precedence and duration when minimizing cycle time. However, disregarding the physical demands of these tasks may contribute to the development of work-related musculoskeletal disorders in the assembly line workers. Three line balancing heuristics that incorporate physical demand criteria were developed to solve the problem of finding assembly line balances that consider both the time and physical demands of the assembly tasks: a ranking heuristic, a combinatorial genetic algorithm, and a problem space genetic algorithm. Each heuristic was tested using 100 assembly line balancing problems. Incorporating physical demands using these algorithms does impact the assembly line configuration. Results indicated that the problem space genetic algorithm was the most adept at finding line balances that minimized cycle time and physical workload placed upon participants. Benefits of using this approach in manufacturing environments are discussed.

Applying Mathematical Modeling to Create Job Rotation Schedules for Minimizing Occupational Noise Exposure

This research developed worker schedules by using administrative controls and a computer programming model to reduce the likelihood of worker hearing loss. By rotating the workers through different jobs during the day it was possible to reduce their exposure to hazardous noise levels. Computer simulations were made based on data collected in a real setting. Worker schedules currently used at the site are compared with proposed worker schedules from the computer simulations. For the worker assignment plans found by the computer model, the authors calculate a significant decrease in time-weighted average (TWA) sound level exposure. The maximum daily dose that any worker is exposed to is reduced by 58.8%, and the maximum TWA value for the workers is reduced by 3.8 dB from the current schedule.

Application of genetic algorithms to the design of lifting tasks

Abstract Proper workplace design for jobs requiring lifting is important in reducing injuries in the workplace. The challenge is to incorporate a variety of postures to meet productivity and safety requirements. In this paper, a form of evolutionary computation, known as a genetic algorithm (GA), is applied to the problem of designing safe lifting tasks within the constraints of the work place. The design criteria is the NIOSH Revised Work Practices Guide for Manual Lifting. The resulting model develops a population of solutions which represent multiple groups of lifting parameters adapted to the productivity and safety requirements set by the ergonomist.

Comparing Statistical and Machine Learning Classifiers: Alternatives for Predictive Modeling in Human Factors Research

Multivariate classification models play an increasingly important role in human factors research. In the past, these models have been based primarily on discriminant analysis and logistic regression. Models developed from machine learning research offer the human factors professional a viable alternative to these traditional statistical classification methods. To illustrate this point, two machine learning approaches - genetic programming and decision tree induction - were used to construct classification models designed to predict whether or not a student truck driver would pass his or her commercial driver license (CDL) examination. The models were developed and validated using the curriculum scores and CDL exam performances of 37 student truck drivers who had completed a 320-hr driver training course. Results indicated that the machine learning classification models were superior to discriminant analysis and logistic regression in terms of predictive accuracy. Actual or potential applications of this research include the creation of models that more accurately predict human performance outcomes.

A genetic algorithm for designing job rotation schedules considering ergonomic constraints

Job rotation is one method that is sometimes used to reduce exposure to strenuous material handling, however, developing effective rotation schedules can be complex in even moderate size facilities. The purpose of this research is to develop methods of incorporating safety criteria into scheduling algorithms to produce job rotation schedules that reduce the potential for injury. Integer programming and a genetic algorithm were used to construct job rotation schedules. Schedules were comprised of lifting tasks whose potential for causing injury was assessed with the Job Severity Index. Each method was used to design four job rotation schedules that met specified safety criteria in a working environment where the object weight, horizontal distance, and repetition rate varied over time. Each rotation was assigned to a specific gender/lifting capacity group. The advantages and limitations of these approaches in developing administrative controls for the prevention of back injury are discussed.

An interactive distributed evolutionary algorithm (IDEA) for design

Evolutionary computation (EC) has been successfully applied to a wide range of design problems. There has also been an abundant amount of work in applying interactive ECs in the design of displays, robot behavior, bitmaps, etc. In the EC literature, one can also see a number of successful design applications of distributed ECs. However, to date, there has been no research in the area of interactive distributed ECs. In this paper, we present an interactive distributed evolutionary algorithm (IDEA) for the design of simple emoticons. We will discuss two ways that our IDEA is currently being used including the areas of EC education and Human Factors.

Interactive evolutionary design of anthropomorphic symbols

Although the computer science literature contains numerous examples that describe various interactive evolutionary computational (IEC) algorithms, few studies have focused on how use such algorithms to elicit design information from a population of human users. The purpose of this study was to address this gap in the literature by constructing and testing an IEC algorithm for anthropomorphic symbol design. A design group of 25 subjects used the algorithm to create 100 anthropomorphic symbols represented the emotions of anger, joy, fear, and sadness. The 100 symbols underwent comprehensions testing using a separate group of 30 subjects. Factor analysis of the nine limb angles comprising each symbol revealed that specific combinations of limb angles differed significantly between symbols based on the emotional referent the IEC algorithm users meant to convey. Comprehension testing results revealed that recognition accuracy for the joy symbols was highest while recognition accuracy for the anger symbols was lowest. The findings of the current study suggest the IEC algorithms can be used to identify important symbol design characteristics and generate symbols whose message is readily comprehended by end user populations.

Using Linear Programming to Optimize Control Panel Design from an Ergonomics Perspective

Linear programming (LP) for optimization of control panel layouts has been incorporating ergonomic constraints into models to reduce reaching distances required for control panel use since the 1960s. These algorithms have used a panels frequency of use, distance from user, and transition distance as basic model variables. A new variation of the LP model for control panel design is proposed that modifies the layout from single point semidry to dual point semidry in terms of design using anthropometrics. The proposed model was applied to the design of a twelve-panel board of six-inch square panels. The model was able to take into account design factors such as control sequence, alignment, and clustering, as well as direct hand access. The resulting control panel solution minimized the reach and movement distances required by an operator. Results suggest that LP optimization can be used to construct “ergonomically designed” control panels that limit MSDs incidence and severity.