Implementing Digital Mammography in your Facility

mplementing FFDM 5-11-0 Implementing Digital Mammography in Your Facility: • PACS • Connectivity • Logistics • Economics Robert J. Pizzutiello, Jr., FACMP Upstate Medical Physics, Inc. Overview • • • • • • • • • Why change anything? The decision process How will FFDM be different? Can we afford it? Workflow Image and data management Mammographers Network considerations Summary Why change anything? Existing technology • Doesn’t meet my needs • Doesn’t meet new regulations • Costs too much • Isn’t “state-of-the-art” • Not compatible with future technology • Doesn’t meet demands of patients or referring physicians The only sure thing in life is change • The cost-benefit analysis • Technological change opens new opportunities Grids, Digital SBB, ID cameras, Mo/Rh – Rh/Rh Extended cycle processing, high-contrast film Daylight loading film processors • Change is not always useful for everyone Xeromammography Dual emulsion mammography film Tilting C-arms (?) Rh and W targets Questions to Consider • What benefits do the new technology provide? • What problems may it introduce? The current, imperfect system works now The new system may require partial – total system change • How will cost (direct and hidden) and revenue change? • Should we change at all? • When should we change? • How do we make the transition? The optimal process for change • Create a team to consider the issue • Create a list of questions to ask? Clinical Technical Regulatory Operational Financial • Assign responsibilities • Obtain proposals from prospective vendors • Site visits, if possible • Consensus decision – with buy-in from all parties • Make a timetable and conversion plan mplementing FFDM 5-11-0 Sub-optimal process of change • Purely financial decisions based only on the cost of goods • Purely political decisions based only on outside pressures • Purely emotional decisions based on a few unfortunate circumstances or personal relationships • Decisions that make sense for general radiology department don’t always work in mammography • Any process that fails to consider the system-wide impact of change Recommended Process • Form the task group Radiologist 8 Clinical QC and Senior staff technologist 8 Technical staff impact 8 Applications and ongoing training Medical physicist and IT specialist 8 Technology assessment 8 Performance specifications Business or office manager 8 Procurement of products, accessories, etc. 8 Office and patient flow issues Purchasing agent or contracts person 8 Protect the facility Digital Mammography • New technology • Regulatory issues Accreditation Timetable QC requirements • Clinical and technological issues Benefits “What will change?” The Imaging Chain Screen-Film vs. FFDM ScreenX-Ray Production …Same Imaging Geometry…Different for Fischer Breast Compression…Different for Fischer Scatter Reduction…Different for Fischer AEC…Different Film Processing…Dry process, less Viewing Conditions…Different “What will change?” Work-Flow WorkSF vs FFDM (DR systems) Patient interview… Same Positioning… Different (Fischer, Lorad) Compression… Different (Fischer) Post exposure… Image appears on monitor Repeat or next view Film Processing… Image transfer Soft-copy interpretation Viewing Conditions… Very important Image Storage… Very different Digital Mammography Image Flow Acquisition 50-60 patients/day Review CAD (Option) Laser Camera (Option) CD-R (Option) Archive (Option) HIS/ RIS (Option) Integrated Digital Solution GE mplementing FFDM 5-11-0 Digital Mammography • 3 facilities considering the change Large, busy private practice Community hospital with satellite centers Large university practice • Review process and decisions • What is motivating this decision to change? Clinical vs. research Early adapters, hi-tech reputation Pressure from patients, referring physicians Limitations of existing S-F system 8 Technical or operational Platform for future development Space Cost Operational Issues • Task group (informally) convened • Busy patient flow • Satellite offices w/o FT radiologist on-site • Image transfer and archive • Desire improved radiologist efficiency • How will we select patients for FFDM? • How much should we advertise this new technology? • Abundant misinformation among patients Space Considerations • Busy practice (9 mammography rooms + …) • Patient volume increasing • Already expanded into lobby, waiting area • No more space in this building • Solution: Digital Mammography increased # patients/rm Added one Spring 2001… and a second in Fall 2001 Operational Issues • Patient selection • Answering patient questions • Training of all staff • No trips to the processor • How to best use RT’s w/o creating burnout? • Additional time and training for QC • Regulatory process • • • • • • • • • • • • Factors to Be Considered Image quality Soft-copy displays (vs. std. PACS monitors) IR size, patient population and selection Clinical vs. research Data transfer and storage Softcopy and hardcopy SF comparison Platform for future development Space and throughput (RT and MD) Changes in staffing, scheduling, waiting room, parking RT training, burnout Space Cost mplementing FFDM 5-11-0 New Imaging - New Quality Control “Can we afford it?” Medicare Physician Fee Schedule 2003 (Non Hospital) Rest of New York 76092 Screening film mammography, bilateral 76090 Diagnostic film mammography, unilateral 76091 Diagnostic film mammography, bilateral G0202 Screening digital mammography, bilateral G0204 Diagnostic digital mammography, bilateral G0206 Diagnostic digital mammography unilateral 76085* Computer aided detection add on code for screening mammography G0236* Computer aided detection add on code for diagnostic mammography $18.04 $78.93 $72.40 $90.03 $125.46 $133.37 $107.40 $18.04 Cases/hour Hours/day Cases/day Days/year 4 8 32 250 Full-Field Digital Mammography Financial Analysis 8,000 $95 85% $646,000 Film cost per year@ $5/case FFDM Capital Cost CAD Capital Cost Years Amoritization Cost FFDM/year Cost CAD/year Annual Net SFM Revenue S/F Annual gross Revenue Film cost per yr.@$5/case Net SFM Revenue $40,000 $400,000 $150,000 5 $80,000 $30,000 $646,000 -$40,000 $606,000 Cases/year Medicare Reimbursement % Operating Capacity S/F Annual Gross Revenue FFDM Multiplier FFDM Gross Annual Income CAD/case CAD Gross Annual Income FFDM + CAD Annual Income 1.5 $969,000 $15 $102,000 $1,071,000 Lorad Net Revenue from FFDM vs. SFM Annual Revenue, FFDM + CAD $1,071,000 Annual Net Revenue, SFM -606,000 Capital Cost (FFDM+CAD), annualized for 5 years -110,000 Annual Increase in Net Revenue for FFDM conversion $355,000 5-year cumulative Increase in Net Revenue for FFDM $1,775,000 conversion mplementing FFDM 5-11-0 University of Oklahoma Digital Mammography Productivity Study From GE web site: http://www.gemedicalsystems.com/rad/xr/mammo/index.html Substantial Film Cost Savings • The Assessment team also found that the digital systems reduced the Institute’s film costs significantly. • Three radiologists read softcopy digital images exclusively, eliminating the cost of film as well as the expenses associated with it. • Result - the Institute is saving at least $128,000 annually on film related costs. Resource Management • Throughput case study • Busy practice • Screening-Diagnostic • 3 SF vs. 2 FFDM machines • 20 - 25 patients/day/machine SF • 30 - 35 patients/day/machine FFDM • 2 FFDM rooms ~ 3 SF rooms 75/day 70/day Human Resource Management • 30 - 35 patients/day/machine FFDM • Initially, 8 hour shifts FFDM • Trials of 4, 2 and 1 hour FFDM shifts • Found 2 hour shifts to be optimum • 1 additional RT handles data entry outside the rooms, supporting 2 RT’s in 2 rooms • Bottom line: 3 RT’s + 2 FFDM ~ 3 RT’s + 3 SF Basic Definitions • PACS – Picture Archiving and Communication System – a Network for image management (capture, storage, distribute and display medical images) • HIS/RIS a (Hospital or Radiology Department) that supports patient registration (or transfer of registration from another system) exam ordering and tracking, film management, and often scheduling, reporting, billing and statistics. mplementing FFDM 5-11-0 MIMS Softcopy WorkStation Images and Patient Information Work List Manager HIS/RIS HL7 To the PACS Broker Retrieve Prior Images DICOM Unprocessed Selenia MIMS Store New Images Pioneer DRM-7000 DVD Jukebox Recommended Process • Convene the technical task group RT, Medical physicist, MD, Manager, IT specialist, FFDM, RIS and PACS representatives Technology assessment Performance specifications (FFDM) 8 Performance specifications (Image management) 8 Obtain commitments 8 Specify in contracts Assure all components work together Produce desired results Site visit to a comparable facility Avoid implementation disasters 8 8 Mammography Image Management Solution Consideration: Image File Size Average Image Size = 22MB (16MB, 28MB) Four Exposures / Study = 88MB Exams / hour = 4 to 6 Ave. Data Collected = 2.8 GB to 4.2 GB /day Annual Storage = 1.1 TB / System Assume Two Priors / Pat. = 12.6 GB over network File Sizes for Digital Medical Images (Megabytes) Modality Mammography Plain Radiographs Fluoroscopy CT MRI US Nuclear Medicine Image Matrix 4096 x 5120 x 12 2048 x 2048 x 12 1024 x 1024 x 8 512 x 512 x 12 256 x 256 x 12 256 x 256 x 8 128 x 128 x 8 Study 4 4 18 30 100 24 24 Images/ File size (Mb) 125.0 25.0 19.0 12.0 10.0 1.6 0.4 mplementing FFDM 5-11-0 Typical File Sizes for Storage Options Radiology CT Image MRI Image CR Image Mammo Image Departmental Archive Text page Novel Library Telephony CD Size 0.5 MB 0.03 MB 7.6 MB 10.3 MB 2 TB 2500 bytes 1 MB 10 TB 22 MB 635 TB Comments (512x512) pixels x 2 bytes/pixel (128x128) pixels x 2 bytes/pixel 14”x17” – (1760x2180) pixels x 2 bytes/pixel 8”x10” – (2032x254(1) pixels x 2 bytes/pixel 100,000 exams/year (17.5% CT, 17.5% MR, 60% CR and 5% Mammography). Uncompressed Single sided, ASCI 400 pages 10 Million Volumes (Library of Congress) One hour recording One hour recording PACS Complexity HIS/RIS PACS Broker (HL7) Modality Acquisition Gateways (DICOM) Database Server Web Server & Clients Archive Server Diagnostic and Clinical Workstations Dual 2K Dual 1K Single 1K Flat Panel Archive Device Adding mammography can double annual storage requirements 1 Terabyte (TB) = 1,000 GB Why Care About the Network? • Critical component of any PACS • Should be one of the first considerations in planning • Frequently “underemphasized” • Very significant consequences result from good (or bad) networking decisions Senographe 2000D Isolated LAN Map Senographe 2000D Isolated LAN Map (Not on the hospital network) (Not on the hospital network) Acquisition Workstation (AWS) (hme1) 192.1.1.100 100 Mb Ethernet (hme0) crossover 192.168.1.10 cable Review Workstation (RWS) Option (hme0) 192.1.1.110 192.168.1.20 100 Mb Ethernet LAN Switch 10/100 Mb 192.1.1.130 192.1.1.120 IDC Optional Archive Device such as “Radstore” DICOM Laser-Cam Senographe 2000D Network Map. Senographe 2000D Network Map. Connected Directly To The Hospital Network. Connected Directly To The Hospital Network. 100 Mb Ethernet (hme0) crossover 192.168.1.10 cable (hme1) Hospital Supplied IP Address (hme0) Hospital Supplied IP Address Optional CD-R and/or Disk Storage Pack (Connected to the Hospital Network via aaGEMS Supplied Router) (Connected to the Hospital Network via GEMS Supplied Router) 100 Mb Ethernet (hme0) crossover 192.168.1.10 cable Senographe 2000D Isolated LAN Map Senographe 2000D Isolated LAN Map Review Workstation (RWS) Option (hme0) 192.1.1.110 Acquisition Workstation (AWS) (hme1) 192.1.1.100 192.168.1.20 SCSI IDC Senographe 2000D Room 100 Mb Ethernet 192.168.1.20 100 Mb Ethernet Optional CD-R and/or Disk Storage Pack LAN Switch 10/100 Mb Using our 3Com switch is recommended but not required. If not used, it can still be useful for troubleshooting. Hospital Network Switch or router 10/100 Mb IDC GEMS LAN Switch 10/100 Mb GEMS supplied Cisco Router Hospital Switch or Router RIS DICOM Laser-Cam Senographe 2000D Room RIS DICOM Archive DICOM Laser-Cam 192.1.1.100 Hospital Network Hospital Supplied IP Addresses and AE Titles Hospital Network DICOM Archive Hospital Supplied IP Addresses and AE Titles mplementing FFDM 5-11-0 Bandwidth • Amount of information that can be sent over a data channel per unit time • Very important consideration given the ever increasing size of diagnostic images • Majority of existing hospital networks inadequate (designed for “text”applications) Bandwidth Examples Analog (modem): Digital ISDN: T-1: T-3: Ethernet: Fast ethernet: OC-3: OC-12: 33-56 kilobits/sec 64-128 kb/sec 1.5 megabits/sec 44 mb/sec 10 mb/sec 100 mb/sec 155 mb/sec 620 mb/sec Archive Future Considerations MIMS Archive Future Considerations ASP Remote Archive 18 months (ideal) of RAID supplied by ASP Vendor LORAD Workflow Manager Storage Area Network • On-line vs. off-line storage • Additional RAID • Faster access • Grows with need • Cost competitive • EMC², IBM, Compaq, DELL • Hospitals may have an Enterprise SAN strategy Off-Site Storage And Disaster Recovery Offsite storage and Disaster Recovery A sample, interesting configuration • ImageCare in Upstate NY Main site – Latham, NY (radiologist on-site) Satellites in Saratoga Springs and Delmar • Goal 3 FFDM systems Cost-effective Centralized interpretation What do we really need? • Main site – the works FFDM with AWS RWS Printer Storage (future) • Satellites FFDM with AWS Data transfer to main site mplementing FFDM 5-11-0 The optimal process for change Satellite FFDM + AWS Satellite 2 FFDM + AWS • Create a team to consider the issue • Create a list of questions to ask? Clinical Technical Regulatory Operational Financial • Assign responsibilities • Obtain proposals from prospective vendors • Site visits, if possible • Consensus decision – with buy-in from all parties • Make a timetable and conversion plan Main site FFDM + AWS Printer Storage (PACS) - future Recommended Process • Convene the technical task group RT, Medical physicist, MD, Manager, IT specialist, FFDM, RIS and PACS representatives Technology assessment 8 Performance specifications (FFDM) 8 Performance specifications (Image management) 8 Obtain commitments 8 Specify in contracts Assure all components work together Produce desired results Site visit to a comparable facility Avoid implementation disasters 8 Technical Details • Transmission method – • Transfer Protocol – • Images transferred (auto push or batch) • Typical transfer time – xx minutes/4-view case • Interim storage solution Soft-copy interpretation Print hardcopy archive until PACS to be installed later this year Take Home Message • Work with your medical physicist and IT people • Write a contract that specifies Quantitative performance parameters Operational performance (example) 8FFDM must interface with PACS and RIS 8The following functions must be supported • Plan in advance Include those who are resistant to change Expect some glitches on first implementation If major problems persist, some will actively resist Summary • • • • • • • • Why change anything? The decision process How will FFDM be different? Can we afford it? Workflow Image and data management Mammographers Network considerations