Comp 15 Usability Human Factors Unit 11 Input
Comp 15 - Usability & Human Factors Unit 11 - Input and Selection Methods This material was developed by Columbia University, funded by the Department of Health and Human Services, Office of the National Coordinator for Health Information Technology under Award Number 1 U 24 OC 000003.
Significance All CIS depend on accuracy and dependability of information entered into them Input methods vary in speed, accuracy, suitability to different users and conditions Technology selection and assessment before deployment is important for success Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 2
Different input methods • • Light pen Pen & touchscreen Mouse Trackball Voice recognition Bar codes, Special-purpose keyboards Device touch pads with LCD screens • Gesture-recognition systems • Haptic control – Usually in conjunction with keyboard » Sittig, 2010 Component 15/Unit 11 • Input selection must be done with: – a view to context (physical, cognitive) – Task – User population – Other variables Health IT Workforce Curriculum 2. 0/Spring 2011 Version 3
Keyboard/Mouse Experienced users up to 15 keystrokes/second (150 words/min) • Beginners <1 keystroke/second (12 words/min) Most common method of input, but requires training, differences in skill level Poor method for healthcare, such as hands-on bedside nursing Excessive use or improper ergonomic conditions can result in to repetitive strain injuries: • Severe injury/permanent loss of hand/arm function • Risk increases above 300 minutes/day • Proper deployment required Paper and pen still used for clinical recording Transfer to other systems --> potential errors, duplication of effort Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 4
Light-Pen Childs: light-pen + user-friendly menus • Users could perform all functions with < 1% typing Pop-up alphanumeric keypads requiring mouse action unsuccessful • Indirect pointing devices require more cognitive processing & hand-eye coordination to bring cursor to target • (from Sittig book) Pop-up digital keyboard may not offer tactile feedback, slow speed of operation, danger of incorrect presses and slow correction Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 5
Voice Input/Speech Recognition Requires adequate hardware and software, including headset microphones or handheld voice recorders Software improves as it is used more and acquires more data about the individual user’s voice Initial use requires a training exercise consisting of reading passages aloud Specialized vocabularies for medicine, specific specialties Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 6
Natural Speech Recognition: Advantages Requires little training except software-specific training Rapid input Handheld recorders; upload for interpretation and transcription Turnaround in hours; user can correct transcript immediately Hands-free Speaker-independent applications good where commands are limited Easy to convert from sending dictation recordings to outside transcription house • In-house speech recognition stations controlled by medical transcriptionists But: interruptions make it less efficient than expected in practice Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 7
Speech Recognition: Disadvantages Requires microphones within certain configuration of face (newer distance desktop microphones are being developed) Misinterpretations by software can lead to uncaught errors Difficult to use in noisy environments; confidentiality may be compromised Software may not fit highly structured forms; compatibility with existing medical records may be a problem Accents or individual speech patterns may require extra training Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 8
Zick & Olsen 2001 • Comparison of voice recognition and traditional transcription service for ED charts • Voice recognition faster (avg 3. 65 min turnaround) v. transcription (39. 6 min) • Transcription more accurate (99. 7% v 98. 5% for voice) • Voice recognition improves in speed and accuracy; decisions made with current technologies at time of deployment Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 9
Tablets • Example: i. Pad • Form factor and light weight; suitable for some applications • 300 medical applications already developed • Meets many requirements for healthcare tablet: – – – Wi. Fi, Dust/liquid resistant Fingerprint authentication Barcode scanning Integrated camera Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 10
Ideal Features for Tablets Ruggedization: • Liquid & crash resistance • Cleanable & disinfectable Swappable battery Gloves should work on it Long battery life Multitasking Barcode scanning Camera (telemedicine, photography of conditions and transmission to consultants) Component 15/Unit 11 Flash to accommodate various websites and video Health IT Workforce Curriculum 2. 0/Spring 2011 Version 11
Touchscreen Input as Part of a Pharmacy Dispensing Unit Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 12
Pen Input • Handwriting recognition: conversion of text in user-drawn image into digital form • Difficult in practice • Gestural alphabets (palmtops, cell phones) differ by brand • Digitizer tablets: learn user’s handwriting; uses samples to train for accuracy • Digital pen and special paper capture and recognize motion sequence • Slower than keyboard input, virtual keyboards replacing it Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 13
Tablets/Pen Input Advantages Disadvantages • Can be efficient in structured formats, lightweight devices, tablets have adequate screen space, easy to learn Component 15/Unit 11 • Screen reading may be difficult in certain lighting conditions, populations (e. g. elderly) • Fragility, weight difficult in bedside practice • Newer, lighter tablets may overcome this • Small PDAs, cell phones have limited screen space, virtual keyboard input methods which may be slow, inaccurate Health IT Workforce Curriculum 2. 0/Spring 2011 Version 14
Pie Menus • Faster than linear menus • Circular menu slices large in size, near pointer for fast interaction (Fitt’s law: ease of target acquisition is proportional to size and inversely proportional to distance) Ideally 3 -12 items; 8 or fewer is best Muscle memory: experienced users need not look Can be nested for many options & pop-up linear menus Shows available options, unlike mouse gestures • • – Depend on direction rather than distance Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 15
Pie Menus (cont. ) • Useful for actions with logical grouping choices • Linear menus useful for dynamic large menus without logical groups • Self-revealing gestural interface • Easy to ‘mark ahead’ because of memory without menu even showing • Eases transition from novice to expert since every use rehearses actions which go into muscle memory & item location unconsciously memorized • Disadvantage: not often available as standard interface widgets (except in games) Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 16
Marking Menus • Similar to pie menus but menu need not appear, multiple actions in chain can select desired item very fast, without need for menu to pop up visibly • Combine pie menus with gestures • See the excellent video at http: //www. markingmenus. org/ Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 17
Contextual Menus (Popup Menus) • Appears on user interaction, in specific context • Limited choices pertaining to current state • Solution to need for rapid selection; also requires little memorization of location • Problems: – Options available only in the context may be confusing & not let user know of availability • Screen edge interactions may be different Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 18
Yen (2005) Digital Pen/Paper • Pen contains camera; records writing pattern on digital paper with 0. 3 mm dots • Camera uses dots to track pen location, creating digital representation in memory • Information transferred to computer; creates digital & paper copies • Study: initial excitement gave way to use interchangeable with regular pens; preference for conventional pens due to bulk, distraction, not fitting in pockets • Nurses saw potential, but physical execution needed improvement Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 19
Gestural Interfaces Examples: • ‘Magic Wall’ election night interface • Wii • i. Phone • no physical contact required; body only input device Need to create discoverable vocabulary, metaphors Physically-oriented applications easy • e. g. Wii for tennis - mapping is directly spatial Fuller use of human body; not just eyes/fingers Pinch to zoom on i. Phone • Direct spatial relationship • Immediate feedback Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 20
Gestural Interfaces (cont. ) • Multitouch 1 point of contact • Gestural commands less obvious, and less obvious than current GUIs • May involve gloves, sensors, multiple cameras, LCD arrays used as pinhole cameras • Output can be of multiple forms: music, video, device control • Examples: http: //www. kickerstudio. com/blog/2009/04/a -collection-of-gestural-interfaces-spring-2009 -edition/ • Interfaces can be almost invisible Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 21
Mechanics: Touch Screens, Gestural Controllers (Saffer, 2009) 1. Sensor 2. Comparator 3. Actuator Component 15/Unit 11 • Detects environmental changes • Size, sensitivity, must be suitable for the purpose, gestures • Usually a microprocessor • Compares current to previous state or goal • Makes decisions • Receives decisions from comparator as a command • Implements action • Physical, e. g, flushing, or • Digital, e. g. changing iphone display orientation when tilted Health IT Workforce Curriculum 2. 0/Spring 2011 Version 22
Gestural Interfaces: Use and Appraisal Consider reason for use? • e. g. public restrooms; people don’t want to touch things, infection control Bad for: • • Heavy/rapid data input (touchscreen keyboards) Visual-only feedback (e. g. for visually impaired users) Reliance on physical (for those with physical limitations Environment (e. g. wearing gloves), subtle movements for those with large hands Contexts where privacy, embarrassment may be issues Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 23
Gestural Interfaces: Advantages Good for: • Natural interactions • Less cumbersome/visible hardware • Flexibility Non-touchscreen gestures: huge number of possible gestures Subtlety, nuance (as with common communication hand facial gestures) Fun, games, play, exploration, teaching Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 24
Gestural Interfaces: Characteristics of Good Design Discoverability Trustworthiness Responsiveness Component 15/Unit 11 • How does one find out the screen is touchable? • How does one discover an environment or device is interactive? • e. g. New York subway has large hand + “Touch, start” on ticket vending machines (attraction affordance) • Must convey lack of danger, privacy • • • Feedback System state, Correct level of sensitivity Aesthetically pleasing Respect for dignity (don’t make users appear foolish in public) Health IT Workforce Curriculum 2. 0/Spring 2011 Version 25
Technology/Input Method Selection • Should be based on consideration of the task, user, and environment • Current technologies (such as voice input) may progress rapidly, so outdated studies and statistics should not be used • Users may have a learning curve; comparison of two different methods should allow for training time (e. g. Kotani: study of pen-tablet v. mouse showed initial mouse superiority; then users became more efficient with pen-tablet • See Mobile Devices for Nursing: a Human Factors Evaluation – http: //www. medicine 20 congress. com/ocs/index. php/m ed/med 2009/paper/view/317 Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 26
Appraisal of Input Methods • Some research on these comes from the field of experimental psychology • Variables measured may include speed of interaction, speed of data entry, accuracy, muscular and cognitive involvement, long-term, short-term and muscle memory, hand-eye coordination • Results can vary from experimental laboratory to in-situ; try to test in the setting in which it will be used, with typical users • Considerations such as size and weight can have significant consequences for whether item will be used, despite sophisticated software or other features Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 27
Questions to Ask in Input Method Selection Does the method require a device? Is fine motor control required? How much training is required? How much practice is required? How much motor intelligence is required? (e. g. typing requires at least weeks of skilled training) Does the user require both hands to operate it? I Is furniture/surface required (e. g. table for keyboard) What speed of data entry is required? Is speed an important consideration? How critical would an inaccuracy be? How easily correctable are mistakes? How easily detectable are mistakes? How many others see the product? (error detection) Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 28
More Questions Does the method have characteristics that make it unacceptable to potential users (e. g. perceptions of rudeness, clerical work… )? Does the input feed into other complex systems (such as a clinical information system, pharmacy order) ? Is structured data required? Is the method going to be used in an ambulatory manner (e. g. by a clinician walking around)? Is disinfection required? To what degree? (e. g. ordinary hospital or clinic v. operating room) Will frequent cleaning/solvents be required? Does the environment require special provisions in setup (e. g. wall or partition separating equipment to protect it from dirt, noise? ) Does the required setup create distance from other clinicians, affecting communication patterns? Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 29
Conclusions • Input methods are an evolving field; standard methods (such as keyboard) will likely persist, but become device-independent • New methods require research, but open up avenues of control useful in medicine. • Gestural non-touch methods may be useful in situations where infection control, freedom of movement, lack of physical devices are key (e. g. ED, OR, pediatrics…) • Matching the input method with senses and modalities involved in other parts of the task is more successful Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 30
Selected References • • • Buxton, William, Issues in Manual Input. User Centered System Design: New Perspectives… (Paperback). by Donald A. Norman, Stephen W. Draper Chapter 15, p 320 -337 Betriebsraum, B. Extremely Efficient Menu Selection: Marking Menus for the Flash Platform. December 11, 2009. Gestural Interfaces. http: //www. kickerstudio. com/blog/2009/04/a-collectionof-gestural-interfaces-spring-2009 -edition/ - A collection of gestural interfaces shown on video http: //www. betriebsraum. de/blog/2009/12/11/extremely-efficient-menuselection-marking-menus-for-the-flash-platform/ Saffer, Dan. Designing Gestural Interfaces Touchscreens and Interactive Devices. . O; Reilly Media. Canada. 2009. See also http: //www. designinggesturalinterfaces. com/ Component 15/Unit 11 Health IT Workforce Curriculum 2. 0/Spring 2011 Version 31
- Slides: 31