Measuring and managing radiologist workload: application of lean and constraint theories and production planning principles to planning radiology services in a major tertiary hospital.

INTRODUCTION: We describe how techniques traditionally used in the manufacturing industry (lean management, the theory of constraints and production planning) can be applied to planning radiology services to reduce the impact of constraints such as limited radiologist hours, and to subsequently reduce delays in accessing imaging and in report turnaround. METHODS: Targets for imaging and reporting were set aligned with clinical needs. Capacity was quantified for each modality and for radiologists and recorded in activity lists. Demand was quantified and forecasting commenced based on historical referral rates. To try and mitigate the impact of radiologists as a constraint, lean management processes were applied to radiologist workflows. A production planning process was implemented. RESULTS: Outpatient waiting times to access imaging steadily decreased. Report turnaround times improved with the percentage of overnight/on-call reports completed by a 1030 target time increased from approximately 30% to 80 to 90%. The percentage of emergency and inpatient reports completed within one hour increased from approximately 15% to approximately 50% with 80 to 90% available within 4 hours. The number of unreported cases on the radiologist work-list at the end of the working day reduced. The average weekly accuracy for demand forecasts for emergency and inpatient CT, MRI and plain film imaging was 91%, 83% and 92% respectively. For outpatient CT, MRI and plain film imaging the accuracy was 60%, 55% and 77% respectively. Reliable routine weekly and medium to longer term service planning is now possible. CONCLUSIONS: Tools from industry can be successfully applied to diagnostic imaging services to improve performance. They allow an accurate understanding of the demands on a service, capacity, and can reliably predict the impact of changes in demand or capacity on service delivery.

Measuring and managing radiologist workload: a method for quantifying radiologist activities and calculating the full-time equivalents required to operate a service.

INTRODUCTION: Accurate and transparent measurement and monitoring of radiologist workload is highly desirable for management of daily workflow in a radiology department, and for informing decisions on department staffing needs. It offers the potential for benchmarking between departments and assessing future national workforce and training requirements. We describe a technique for quantifying, with minimum subjectivity, all the work carried out by radiologists in a tertiary department. METHODS: Six broad categories of clinical activities contributing to radiologist workload were identified: reporting, procedures, trainee supervision, clinical conferences and teaching, informal case discussions, and administration related to referral forms. Time required for reporting was measured using data from the radiology information system. Other activities were measured by observation and timing by observers, and based on these results and extensive consultation, the time requirements and frequency of each activity was agreed on. An activity list was created to record this information and to calculate the total clinical hours required to meet the demand for radiologist services. RESULTS: Diagnostic reporting accounted for approximately 35% of radiologist clinical time; procedures, 23%; trainee supervision, 15%; conferences and tutorials, 14%; informal case discussions, 10%; and referral-related administration, 3%. The derived data have been proven reliable for workload planning over the past 3 years. CONCLUSIONS: A transparent and robust method of measuring radiologists' workload has been developed, with subjective assessments kept to a minimum. The technique has value for daily workload and longer term planning. It could be adapted for widespread use.

Measuring and managing radiologist workload: measuring radiologist reporting times using data from a Radiology Information System.

INTRODUCTION: Historically, there has been no objective method of measuring the time required for radiologists to produce reports during normal work. We have created a technique for semi-automated measurement of radiologist reporting time, and through it produced a robust set of absolute time requirements and relative value units for consultant reporting of diagnostic examinations in our hospital. METHODS: A large sample of reporting times, recorded automatically by the Radiology Information System (COMRAD, Software Innovations, Christchurch, New Zealand) along with the description of each examination being reported, was placed in a database. Analysis was confined to diagnostic reporting by consultant radiologists. A spreadsheet was produced, listing the total number and the frequency of reporting times of each distinct examination. Outliers with exceptionally long report times (more than 10 min for plain radiography, 30 min for ultrasound, or 60 min for CT or MRI with some exceptions) were culled; this removed 9.5% of the total. Complex CTs requiring separate workstation time were assigned times by consensus. The median time for the remainder of each sample was the assigned absolute reporting time in minutes and seconds. Relative value units were calculated using the reporting time for a single view department chest X-ray of 1 min 38 s including verifying a report made using speech recognition software. RESULTS: A schedule of absolute and relative values, based on over 179 000 reports, forms Table 2 of this paper. CONCLUSIONS: The technique provides a schedule of reporting times with reduced subjective input, which is more robust than existing systems for measuring reporting time.

The use of visual search for knowledge gathering in image decision support.

This paper presents a new method of knowledge gathering for decision support in image understanding based on information extracted from the dynamics of saccadic eye movements. The framework involves the construction of a generic image feature extraction library, from which the feature extractors that are most relevant to the visual assessment by domain experts are determined automatically through factor analysis. The dynamics of the visual search are analyzed by using the Markov model for providing training information to novices on how and where to look for image features. The validity of the framework has been evaluated in a clinical scenario whereby the pulmonary vascular distribution on Computed Tomography images was assessed by experienced radiologists as a potential indicator of heart failure. The performance of the system has been demonstrated by training four novices to follow the visual assessment behavior of two experienced observers. In all cases, the accuracy of the students improved from near random decision making (33%) to accuracies ranging from 50% to 68%.