Experience with implementation of a radiology speech recognition system

Recent advances in speech recognition technology have allowed development of computer systems for real-time radiologist-driven generation of reports. The transition to a speech recognition system is a technically complex process with many potential pitfalls that can decrease efficiency and disrupt workflow. In our recent experience with installation of such a system in an academic radiology department, factors that have worked against optimal performance have included environmental logistics, hardware incompatibilities, radiology information system interface problems, lack of suitable training, and inadequate technical support. Communication of our experience is intended to allow radiologists to anticipate complications of these systems and make informed decisions regarding the feasibility of such a system in their practices. With this information, potential buyers should be able to carefully scrutinize specifications for prospective systems and, by avoiding many of the possible pitfalls, make an easier transition to a speech recognition environment.

Injury in traditional and sport rock climbing

The objective of this study was to compare patterns of injury found in traditional rock climbing with those found in sport climbing. A questionnaire was administered to rock climbers by mail, in person, and via the World Wide Web. Injuries that occurred while rope-protected climbing on rock were analyzed regarding the anatomical location and the mechanism and activity at the time of injury. Ninety-four climbers reported sustaining an injury while rope-protected climbing on rock. Most injuries occurred while leading and involved the upper extremity, especially the fingers. Falling was the predominant mechanism of injury on traditional climbs, and stress over a joint while attempting a difficult move was the most common mechanism on sport climbs. Potential for injury prevention lies in teaching climbers to recognize the limitations of the fingers as weight-bearing structures.

Injury in traditional and sport rock climbing

The objective of this study was to compare patterns of injury found in traditional rock climbing with those found in sport climbing. A questionnaire was administered to rock climbers by mail, in person, and via the World Wide Web. Injuries that occurred while rope-protected climbing on rock were analyzed regarding the anatomical location and the mechanism and activity at the time of injury. Ninety-four climbers reported sustaining an injury while rope-protected climbing on rock. Most injuries occurred while leading and involved the upper extremity, especially the fingers. Falling was the predominant mechanism of injury on traditional climbs, and stress over a joint while attempting a difficult move was the most common mechanism on sport climbs. Potential for injury prevention lies in teaching climbers to recognize the limitations of the fingers as weight-bearing structures.

Injuries in whitewater kayaking

OBJECTIVE

To provide epidemiological data on whitewater kayaking injuries using a descriptive study.

METHODS

A retrospective survey was distributed at whitewater events and club meetings, and made available and advertised on the world wide web, through postings and announcements to newsgroups, related sites, and search engines. Data on sex, age, experience, and ability were collected. Injury data collected included mechanism, activity, difficulty of rapid, and self reported severity.

RESULTS

Of the 392 kayaking respondents included in the final analysis, 219 suffered 282 distinct injury events. The number of days spent kayaking per season was the only independent predictor of injury. The overwhelming majority of injuries occurred while the kayaker was still in the boat (87%). Striking an object was the most common mechanism of injury (44%), followed by traumatic stress and overuse (25% each). The most common types of injury were abrasion (25%), tendinitis (25%), contusion (22%), and dislocation (17%). The upper extremity, especially the shoulder, was the most commonly injured area of the body. Although half of injured kayakers sought medical care for their injury, and almost one third missed more than one month of kayaking because of the injury, almost all (96%) reported a complete or good recovery.

CONCLUSIONS

Factors relating to likelihood of injury appear to be connected with exposure, namely the number of days a year that the sport was pursued. Except for class V (extreme) kayakers, reports of injuries paralleled the number of participants. Kayakers reported injuries predominantly on rivers that they assessed to be at a level appropriate to their skills.

Computer-based collection of mammographic exposure data for quality assurance and dosimetry

PURPOSE

There is potentially more to quality assurance in mammography than the MQSA mandated tests. In this paper we describe a method of capturing individual mammogram technical parameters and the creation of new measures. These include the numbers of images required for each screening examination by technologist, median compression by technologist, and the radiation dose of the examination to the general population of patients.

METHOD/MATERIALS

With this method we describe a semiautomated method of the collection of technical data from mammography exposures. The data that are automatically created by the mammography unit are saved on a computer for later analysis. The method was used on 2738 consecutive screening mammography examinations and 13 621 exposures from one machine. Data were obtained from November 1998 through December 1999.

RESULTS

Using standard methods, the mean glandular dose (MGD) per exposure was 2.62 mGy (SD 1.2). The mean dose per bilateral screening examination was 6.53 mGy (SD 3.07), the median dose was 6.11 mGy, and the dose range was 1.13-34.23 mGy. Rhodium filtration was used for 18% of the exposures. The average and median breast thickness was 4.9 cm. The ACR phantom MGD for this machine was 2.44 mGy at 25 kVp, and 1.97 mGy at 26 kVp. The mean number of exposures for a bilateral mammogram was 4.9, and varied by a technologist from 4.7 to 5.2. The mean compression pressure varied by technologist from 13 to 30 lbs (58-134 N).

CONCLUSIONS

The mean dose per mammogram is slightly greater than the ACR phantom dose at 25 kVp. Almost five exposures were necessary for a standard bilateral examination, and this varied by technologist. The compression used also varied by technologist. The semiautomated collection of technical data can aid in maintaining an effective mammography QA program.

Online medical surveys: using the Internet as a research tool

Technologic advances and popularization of the World Wide Web now allow researchers to use the Internet for medical surveys. The Internet provides access to a rapidly growing and widespread population of potential research subjects. Specific populations can be selected from the larger Internet community, which is itself becoming more representative of society as a whole. The hypertext markup language can be used to create interactive forms with an intuitive and intelligent interface. Response collection by means of common gateway interface scripts can afford automated data compilation and exportation to other software packages. Furthermore, real-time evaluation of incoming data can ensure that the data are complete and accurate before they are accepted. Used properly and for appropriate research projects, the Internet can provide a fast, cost-effective, and efficient mechanism for administering medical surveys.

Extinction coefficient determination using target reflectance measurements

Laboratory measurements are reported for optical extinction at a wavelength of 1.06 microm in water droplet clouds. The extinction coefficient, sigma(T), is determined using the two-way attenuation of a target reflected signal and comparing it to the extinction coefficient sigma determined by a single-pass transmission measurement. As well as solid targets, layers of the clouds have been used as a reflector by employing a selective chopping method to provide range-resolved backscattering information and replicate in the laboratory a lidar configu-ration. It is found that multiple scattering can lead to substantial differences between sigma(T) and sigma and that these differences depend upon the properties of the scattering medium and the target as well as on the field of view of the backscatter receiver used for the reflectance measurements. By keeping the field of view very small, the two methods of measuring the extinction coefficient give the same values.

Transmitted beam profiles, integrated backscatter, and rangeresolved backscatter in inhomogeneous laboratory water droplet clouds

Using laser sources at wavelengths of 1.06 and 10.6 microm, transmitted beam profiles, integrated backscatter, and range-resolved backscatter were measured in laboratory-generated water droplet clouds. Clouds with carefully controlled properties were produced in a specially designed cloud chamber. Inhomogeneities were introduced by partitioning the cloud chamber into three adjacent sections separated by air screens. The measurements show the influence of multiple-scattering effects in both the forward and backward measurement geometries, and these are investigated as functions of optical depth, cloud inhomogeneity, and receiver field of view. These data are unique in many ways, and they provide a great deal of insight to the scattering processes which directly affect lidar-type measurements. As well, these measurements provide a welldocumented and detailed database for model validation. Very good agreement is demonstrated with the solutions derived from the multiscattering propagation model described in a companion paper [Appl. Opt. 27, 2478 (1988), same issue].

Four-component polarization measurement of lidar atmospheric scattering

Measurements of the backscattering from the lower atmosphere of linearly polarized ruby lidar pulses at 694 nm are described. A four-channel receiver provides simultaneous measurements of the four components of the Stokes vector of the scattered radiation. These data are used to obtain information on the polarization state of the scattering and its dependence on atmospheric conditions. Results are presented and discussed for clear air, hazes, and several cloud types. In water droplet clouds some rotation of the plane of polarization and changes in ellipticity are noted in the backscattered wave. Linear polarization of the penetrating lidar beam is strongly preserved even in quite dense clouds. In view of the increased complexity of the four-component measurement technique it would appear that in such cases two-channel measurements of delta would be the most efficient lidar approach.