Medical Human Factors Exam 1

What can go wrong with infusion pumps?

- Programmable for use in multiple scenarios (continuous infusion, intermittent, patient-controlled, etc.) which can be mixed up - Alarms that beep at too many things will be ignored or even shut off if they aren't meaningful; lots of things beep in a hospital - Ex. of clip stopping flow but machine says it's going because it's pumping - Ex. of warning labels on side which won't be seen - Baxter's recalled pump could shut down accidentally while delivering medication; confusion between start and on/off button - Maybe can't see screen in operating theater - Making one entry error causes user to turn it off and back on - With smart pumps that set limits on delivery based on drug and patient characteristics, can still go wrong if wrong patient/drug/weight/etc. entered

Science of Human Factors: Separating Fact from Fiction

1. HF is about designing systems that are resilient to unanticipated events, not about eliminating human error 2. HF addresses problems by modifying design of system to better aid people, not by teaching people to modify behavior 3. HF work ranges from individual to organizational level 4. HF is a scientific discipline that requires years of training and often extra degrees 5. HF personnel represent different specialty areas and methodological skill sets

Norman's error classification system

1. capture - frequently done action takes over correct action 2. description - correct action applied to wrong object 3. loss of activation - forgetting something in the action sequence 4. mode - same action does different things depending on mode of system 5. data driven - external data triggers automatic response that is inappropriate 6. associative activation - same as data driven, only source of data is internal

Todd - Injury and Death Associated with Hospital Bed Side-Rails: report to the FDA

643 events (72 fatalities) recorded over a 10 year study span - #s are likely to be low due to reporting deficiencies. Authors recommend inspection of beds for possible entrapment zones, using compatible side rails, verifying installment of side rails, and using supplementary safety measures like checking on the most at-risk patients. What should really be done: change design so that entrapment is not possible.

Vicente - Programming errors contribute to death from patient-controlled analgesia: case report and estimate of probability

A 19 year old woman emerged from a C-section healthy but died of a morphine overdose. The standard 1mg/mL solution was unavailable, so the nrse used a cassette of 5 mg/mL but incorrectly entered the concentration in the infusion pump.

Give examples of walk up and use devices. What are some of the challenges? What are the challenges of designing for global health?

AEDs, Glucose meters, home blood pressure cuffs, diagnostic kits for cholesterol/UTI/HIV/prostate cancer/drugs/pregnancy. For home diagnostic kits, have to ensure that kit is constructed in a way that precludes improper sample collection, presents the results in a clear way, and provides medical contact for results of an urgent nature (HIV testing). For AEDs, must be used quickly, correctly, require no training, and must not be used in non-critical situations. For global health: must be language neutral, graphics must be culturally appropriate, must be maintainable in situ using local materials, used with limited power, truly walk up and use, have immediate observable benefits to user, and strongly mitigate improper uses.

Accuracy, specificity, and sensitivity

Accuracy: (true positives + true negatives)/all cases Specificity: proportion of negatives identified as negatives (high specificity = low Type I error) Sensitivity: proportion of positives that are identified as positives (high sensitivity = low Type II error)

What is human error?

Actions that reduce, or have the potential to reduce the effectiveness, safety, or performance of a system. Must relate to the system at hand but actual adverse consequences not required. Reason's (1992) definition of error: A failure to achieve the intended outcome in a planned sequence of mental or physical activities when that failure is not due to chance (and it is almost never just chance).

McArthur - Submaximal decision theory and health resource conservation: the example of congestive heart failure

Addresses the challenge of the complexity and variety of chronic illnesses in the US and of how those suffering from them react to different treatments. The article suggests using a "personal normal" derived through patient's own clinical history analyzed by info-gap, and distributed health care such as wireless biosensors and community health workers that would allow greater self-management.

Why use surgical robots? What kinds are there?

Allows for minimally invasive surgery; provide better control than typical laparoscopic methods; surgeons can use more than two hands; filters tremor; movement scaling; allow magnification of surgical site for greater precision; enhances human capabilities Types: Supervisory controlled - surgeon must map out entire surgery beforehand, robot performs surgery according to the specific instructions, no midcourse corrections; have to worry about surgeon's ability to take control, mapping has to be perfect, and landmark points have to be perfect Robotic drilling robots Tele-surgical robots - surgeon directs robot's movements, can switch tools, can have multiple arms working simultaneously; surgeon is typically not working directly over patient Shared control systems - system aids surgeon but surgeon must still perform each step and is typically over the patient Steady Hand Da Vinci - very precise; some haptic feedback; focused on one specific spot so can't see what's happening around the site of surgery; no nonverbal communication b/w surgeons or with crew; no smells conveyed to surgeon; could load wrong tool; costs about $2M with high maintenance costs

Sandra Robinson - Consumer Views of Hospital Errors in the US

Americans more likely in 2000 than 1996 to say there are big differences in quality of health care providers; majority found information comparing medical errors useful for determining quality; Americans more concerned about safety in health care than in travel or buying food; favor mandatory reporting and public availability of medical error information; starting to believe they are partially responsible for ensuring safety; preferred tips for safer health care that were directive and specific (make sure dr. knows your medications) vs. ones seen as embarrassing or rude (ask dr. if he washed hands) 42% of Americans believe they have experienced a medical error

Nelson - Use of SoloShot autodestruct syringes compared with disposable syringes in a national immunization campaign in Indonesia

Autodestruct syringes can reduce improper reuse of syringes, which presents a significant risk of infection, especially in developing countries. SoloShot is an auto-destruct syringe which has been shown to be safer and easier to use than standard syringes. Observation and dose measurements revealed that SoloShot delivered more precise and consistent doses and 15% more doses per vial than disposable syringes. Vaccine savings may be partially offset by higher price of SoloShot, but vaccinators preferred it, describing it as easy to use, faster, and more accurate.

Houtchens - Medical-care systems for long-duration space missions

Because of the inability to return to Earth rapidly and conveniently, the capability of delivering medical care on site will be key to success of a manned space station, lunar base, and Mars mission. Spaceflight medical equipment must meet constraints of size, weight, power requirements, and must function accurately in remote, self-contained, microgravity settings after extended intervals of storage, with neither expert operators nor repair technicians on site. Satisfying these requirements will require validation in clinical settings as well as in simulated operational settings. The paper distinguishes between 3 classes of medical issues, and recommends focusing on Class 2, those problems for which the presence of an inflight medical system would have the greatest impact on clinical outcome and mission success.

The Embryo Imbroglio

Before the eggs were implanted, an embryologist realized he had prepared a sample from another woman but decided not to say/do anything because of low chances of viability. How did this happen? - didn't keep eggs as separate as he should have, wasn't following procedure - didn't want to admit that he made a mistake or make patient nervous that he would mess up again - in future, could have a checklist for verifying eggs before loading them, but it was pretty good system overall

Nodine - using eye movements to study visual search and improve tumor detection

Can use eye tracking as a perceptual aid by looking at areas where physicians dwelled and refocusing on those areas with circles. Issues: intrusive, calibration issues, multi image issues, fatiguing for reader

Aylward - Reducing the risk of unsafe injections in immunization programs: financial and operational implications of various injection technologies

Compared 4 types of equipment used to administer vaccines: 1. sterilizable needles and syringes 2. autodestruct needles and syringes 3. standard disposable needles and syringes 4. jet injectors. Owing to differences in cost and other factors, in some settings a combination of equipment is best. For example, autodestruct syringes may be used in areas where it is difficult to ensure adequate supervision, while in medium-sized, fixed-site clinics with safe injection practices, sterilizable equipment will be the most cost effective.

Cooper - Preventable Anesthesia Mishaps

Critical incident analysis on human error and equipment failure in anesthetic practice showed that most preventable mishaps involved human error (82%), and while overt equipment failure accounted for 14% of errors, poor equipment design was also involved in many human error incidents. Other common factors were lack of communication, haste or lack of precaution, and distraction. Mostly training issues (inability to use equipment or tell when there's a problem) and not really a problem with vigilance while monitoring patient. Issues with using interviews: people might not admit everything; interviewer can unintentionally introduce bias; events that were big or happened on days with unrelated big events stand out more than small, common events or events on normal days

Krupinski - the importance of perception research in medical imaging

Current research suggests that there is a false negative rate of 20-30% in radiology and a 2-15% false positive rate. Understanding fundamentals of visual perception can improve delivery of medical care by facilitating better diagnoses through enhanced automation, detection, and strategies for identification. Expert radiologists are not any better at finding Waldo, but they look much longer, suggesting that perceptual search strategies are not driving their abilities. CAD systems can improve diagnostic outcomes by focusing attention on potential lesion areas, suggesting second looks for highly suspect areas, and decreasing distractors. Can worsen diagnostic outcome by making dr. focus only on regions identified by computer, become reliant on algorithm, or increase false positives if computer identified.

What is user centered design? What are some challenges of UCD in MHF? How are they addressed?

Design focusing on and involving the user throughout; special focus on understanding users needs, requirements, and use patterns before designs are initialized. With MFH: hard to interrupt or get feedback during certain procedures such as surgery or emergency medicine; can't put prototypes into the system; difficult to account for all the contexts of use for some medical devices (off label use can be rampant and undocumented) Addressing concerns: prolonged data collection efforts; focus on non-field data collection methods (critical incident analysis, cognitive walk throughs); simulators

Kossack - Industrial Design and Human Factors: Design Synergy for Medical Devices

Design has an impact on both perceived and actual usability of devices. A collaboration between industrial designers and human factors engineers can improve both usability and elegance. Emotional design should be considered, which includes visceral processing (eliciting appropriate psychological or emotional response), behavioral processing (complementing a user's own behaviors, assumptions, and mental models), and reflective processing (reflecting on one's overall impression of a product).

How did Veronica define design? What are constraints of low resource settings?

Design: the iterative process of defining the problem, generating ideas, evaluating those ideas, constructing prototypes, and testing. Does NOT have to be complicated. Constraints: electricity and clean water not reliably available, no or minimal lab equipment, concentration of technology in cities, reluctance to change or learn new procedures, fragmentation of the developing world (lack of organized medical systems), brain drain (no incentive to invest in training), cost, lack of spare parts and repair technicians

Field - Safe Medical Devices for Children

Designing for children: size scaling (things need to be smaller); reduce dosage; growth considerations (ex. put in new pace maker as child grows?); physiologic differences (ex. account for ability of body to handle pressure). Other considerations: not all children are same size, child devices can be very expensive, need special devices to immobilize children for treatment they won't like.

Obradovich - users as designers: how people cope with poor HCI design in computer-based medical devices

Device was retrofitted to do standard IVP, originally used for insulin administration. Really hard to prevent errors caused by a device that has been modified for a new purpose, but people do that because it's cheaper and already available. Big problem in this case: with the modal system, people wouldn't recognize which mode they were in and would make mistakes. Modal nature leads to errors in mode selection, errors in determining location in programming sequence, errors in interpretation of data presented on a display, sometimes people just keep pressing buttons until something happens. Users generated adaptive behavior to account for poor interface: rewriting the guide, developing new procedures, eliminating old ones, implementing memory aids. Even if they "fix" the problem, these adaptive behaviors are bad because they may not be very fault tolerant and induced errors in new procedures may be severe. Lesson: necessary to test devices using error analysis to see what happens when users fall off the path (do they make more errors or see them and fix them?); users should NEVER have to modify devices/procedures to accomplish goals

Patient informatics

Discusses 4 aspects of patient informatics: 1. Internet for patient medical education 2. Web portals for patient access to healthcare systems - web based programs that allow access to health related services (online registration, medication refills, lab results, e-visits, patient education, storage of personal health records, appointments, bill paying, logs for vital signs and tracking of health data) 3. Personal health records - electronic, universally available, lifelong resource of health data; likely a few years in the future 4. Patient-physician communication via email and e-visits

Problems with online diagnostic tools

Don't have option to check multiple symptoms; have to choose from a list of symptoms which may be incomplete; many symptoms are generically associated with a wide variety of diseases; often less helpful than just googling; for some problems, like burns, there is too much science and not enough descriptions of each kind of burn or when to go to the doctor

Give an example of the devices Veronica has overseen.

Dose-Right clip - HIV infected children require liquid medication, but in US 40-60% of caregivers make mistakes when administering it. Solution: clip which costs 8c/unit and solves problem by preventing plunger from going back too far. CPAP - addressing problem of infant respiratory distress. Used an aquarium pump (low cost and durable), got feedback in Malawi and made changes based on it, improved survival rate from 0% with no treatment and 25% with oxygen to 70% with the CPAP Breath alert - apnea detection and correction monitor. Stretch sensor around baby's stomach activates alarm light and vibrational motors if baby stops breathing, wakes it up. Great help to overburdened nurses, who don't have to spend time monitoring baby.

Cook - safety technology: solutions or experiments?

Emphasis on technology in improving safety is pervasive; problem with tech applications is the difficulty in discovering and addressing undesirable complications and side effects; nurses play an integral role in this process. Ex. with a pick list for prescriptions, easy to hit the wrong one; bar code systems where nurses scan a duplicate and get around to actually dispensing medication when they have time - increases chance of misdelivery

Martin - Medical Device Development: the Challenge for Ergonomics

Ergonomics methods: Contextual inquiry - a semi-structured interview method to obtain info about context of use, where users are observed and asked questions while they work Cognitive task analysis - mapping a task, identifying critical decision points, etc. Heuristics - using principles that generally make for more usable/innate designs Usability testing - self-explanatory Cognitive walkthroughs - learning through exploration by having an expert evaluator perform various tasks

Do It By Design

Examples of design errors: - Installation of IV pump cassette from another model - Patient death due to oxygen flow meter being installed upside down - Patient death because alarm couldn't be heard over machines own operating noise level - Surgical equipment packaging prevents sterile delivery

Johnson - The Causes of Human Error in Medicine

Factors that cause error arranged in levels: Level 1 - factors that influence behavior of individual clinician (poor equipment design, technical complexity, multiple competing tasks) Level 2 - factors that affect team performance (problems coordinating or communicating, acceptance of inappropriate norms, development of coping strategies) Level 3 - factors that relate to management of health care applications (poor safety culture, inadequate funding and resources, inadequate risk assessment and clinical audit) Level 4 - involvement of regulatory and gov'tal organizations (lack of national structures supporting information exchange and risk management, preservation of local and professional autonomy against national standards for reporting and analyzing error, focus on device failure and not on device use) Intentions behind patient safety: altruism, media interest, social trends and public opinion, insurance and litigation (finance)

IOM - To Err is Human: Building a Safer Health System

First mention 98,000 figure (# people who die from medical errors in hospitals), making it a leading cause of death How to design a safer health system that acknowledges the tendency of people to make mistakes. IOM expected a 50% reduction over 5 years. Suggestions: 1. establish national focus to enhance knowledge base about safety and develop evidence-based understanding of errors, 2. raise standards and national goals for improvements in safety, 3. implement safe practices at delivery level, 4. identify and learn from errors through voluntary and mandatory reporting practices

Sokol - a needs assessment for patient safety education: focusing on the nursing perspective

Focus group of nurses addressed issues related to reducing health care errors. Biggest obstacles included the system and culture of tolerance (it's ok to commit an error; culture includes code of silence; reactive rather than preventative or educational policy), barriers to reporting and resolving error (individual may not want to report because of fear of punishment/embarrassment), education and training. Suggestions: report near misses, learn from mistakes, utilize error support groups, use nurse market leverage to force employers to create better safety policy and error reporting.

Schwartzberg - Low health literacy: what do your patients really understand?

Health literacy is the ability to evaluate health information for credibility and quality, analyze relative risks and benefits of treatment, calculate dosages of medication, interpret test results, and locate health information. Low health literacy: 6% more hospital visits, 2 days longer in hospital, fewer dr. visits but higher hospital costs, estimated $73B additional health care expenditures, 2x as likely to report poor health/be hospitalized. What to do: pictorial/graphics only, multi-language support, write at 5th grade reading level or below, simplified step by step instructions, anticipate and resolve issues before fielding, read and review instructions with patient.

Why is error classification important? What are some of the pitfalls?

Helps with finding patterns, seeing which errors are most common or most serious, comparing and collecting mass data. Pitfalls: when one error could fall under multiple headings, creating a list so long that people just memorize a few categories (like Gawron's) Solutions: make like google and describe actual error instead of bucketizing it

What is the current mission scope for the ISS? What are some of the operational challenges? What are considerations for future exploration missions?

Mission scope: 3-6 astronauts staying 6 months; physician not required to fly on every ISS mission; astronauts pass rigorous health tests; crew medical officers are remotely guided by flight surgeons in mission control; on-orbit drills and refresher training for things like CPR Operational challenges: zero gravity; limited ground training of CMO; ruggedized equipment and electronics; backup paper copies of emergency procedures only; loss of signal communications with mission control; up-mass and storage limitations; international partner considerations Future considerations: further out we go, less possible a return to Earth; vehicle and hab size depend on mission; increased time delays for communication; maintenance/repair of equipment; what kind of physician should be on board (emergency doc, general, specialist, psychiatrist)?

Leape and Berwick - 5 Years After to Err is Human

NO SIGNIFICANT IMPROVEMENT AFTER 5 YEARS. But, Agency for Healthcare Research and Quality (AHRQ) and $50M/year for research in patient safety as a result of report. Progress has been slow but IOM report brought error prevention to light as a worth goal, changed conversation to focus on systems, stimulated stakeholder engagement in patient safety, and motivated hospitals to adopt new safety practices. Barriers to further progress: the culture of medicine (which emphasizes individual performance and autonomy, not very team-like, deference to surgeon), complexity of health care technology and environment, physician fear of lawsuits, difficulty in measuring progress, current structure of payment rewarding error (more billable business), fragmentation and specialization of care

Safer than Safe - the Polio Vaccine

No fraud or greed involved, but failed safety checks allowed for the deaths of a few hundred people. 2 problems: the testing (statistical instead of each batch; sensitivity too low) and the scale-up (instruction manual didn't include times between steps, making bigger batches complicated the process more than realized, was not treated as a new process that needed oversight)

Qualities of a good reporting system

Non-punitive, confidential, independent, evaluation by experts, timely, system focused, responsive. Issues with self-report: captures only about 30% of incidents (MULTIPLY BY 3.5 FOR REAL NUMBER)

What are the legal limits on sleep deprivation for doctors and nurses? Why the long hours?

Nurses: no national standard; states with regulations vary between 8-12 hours of continuous duty time, with mandatory rest cycles of 8 hours and exceptions for emergencies - for pilots, federal gov't issues licenses; for doctors and nurses, the state issues them Doctors: maximum duty period of 24 hours, with additional 6 allowed for education and transfer of patients; max 80 hour work week averaged over 4 weeks; guidelines are voluntary and 83% of interns reported working more hours than allowed by the standard Why: tradition/rite of passage, cost containment (three shift out of one intern vs. one shift each for 3), medical necessity (avoiding handoffs, seeing case through to completion)

What is the problem with using the internet as a "trusted source"?

Official pages (gov't entities, professional organizations, WebMD) aren't always the first results. It's hard to tell which sources can be trusted because unlike print sources, which must be peer reviewed and require a large monetary investment, all publishers compete on the same access field with the internet and there is no filter for quality. Bad information can be well-written, cite many sources, and look professional. As we saw with internet diagnosis exercise, it's really difficult to get the right diagnosis online. As we saw with the informatics exercise, even highly internet-savvy users have trouble discerning quality of information (range of ratings from 20-100%; huge disagreement b/w websites; some very bad studies cited).

Graber - diagnostic error in internal medicine

Out of 100 cases of diagnostic errors, 90% of cases resulted in harm to patient. Many errors are co-varying, and early mistakes tend to lead to more mistakes. 3 kinds of diagnostic errors: 1. No fault errors (masked or unusual presentation of disease, patient related errors) 2. System related (technical failures and equipment problems, organizational flaws) 3. Cognitive (faulty knowledge, faulty data gathering, faulty synthesis) Issues with this paper: Should no fault be a category? For deceptive patient, could do tests to corroborate claims; dr. should konw if a patient has dementia or other memory problems Close call errors were scored as zero; doesn't account for under-reporting from doctors; psychological impacts scored as zero

Cermack - Monitoring and telemedicine support in remote environments and in human space flight

Paper explores criteria for the selection of portable telemedicine terminals in remote terrestrial places, characteristics of currently available mobile telecom systems, and the concept of integrated monitoring of physiological and environmental parameters. It also describes aspects of emergency medical support in human orbital spaceflight, the limits of telemedicine support in near-Earth space environment, and mentions some open issues related to long-term exploratory missions beyond low Earth orbit.

Set Phasers on Stun

Patient goes in for standard radiation treatment, is subjected to significant amounts of over-radiation, and dies. Factors: - Lack of feedback b/w patient and nurse and machine and nurse - Equipment issues - camera and mic trained on patient not working (if it had, he would have gotten only one shock) - Instead of halting everything until system was running right, they proceeded anyway - Radiation machine issues - did not convert right from one mode to the other; didn't tell nurse it had delivered radiation; error code instead of words or a forced shutdown Machine failed because it didn't expect anyone to make a mistake in entering mode, couldn't handle such a quick correction, combined two very different modes in one machine, and didn't separate the modes well enough

Double Vision

Physician diagnosed an anomaly in a leg x-ray as a rare tropical parasite and made plans to operate, when really it was a result of cleaning fluid congealing on the x-ray machine. How did it get there? Should redesign machine so that cleaning fluid won't get trapped. Leader exerting undue influence on medical students, who saw error but were afraid to speak up. How to get people to speak up to superiors? Flag that all the diseases she hypothesized were rare tropical issues, since patient hadn't been exposed to those areas of the world. Physician was searching for a solution and didn't think of causes coming from information chain, just looked for explanations to fit the x-ray image. Then told another dr. her hypothesis. Should ask a blank question to allow for a non-biased opinion.

Krupinski - perceptual enhancement of tumor targets in chest x-ray images

Precuing the location of potential lung tumors is less effective than precuing and bounding the region of interest with a circle. Circling the ROI was found to focus eye fixations and increase accuracy of finding the tumor. When targets were placed outside the circle, their detection was inhibited relative to targets within the circle. These findings suggest that cuing by circling restricts target detection to the ROI and reduces noise.

Zhang - Wearable medical devices for tele-home healthcare

Presents an overview of wearable technologies for remote patient monitoring. Current project aims to develop a system which utilizes small, wearable biosensors for continuous acquisition of multiple biosignals, short-range wireless communication b/w sensors and a home-based intermediate terminal, mobile phone communication for remote data access, and multisensor data fusion methods to provide pre-diagnosis. They would integrate a cuffless blood pressure meter, finger ring sensor for heart rate monitoring, and bluetooth based ECG monitor.

Will you be misdiagnosed? How diagnostic errors happen

Rate of misdiagnosis ranges from <5% in perceptual specialties (pathology, radiology, dermatology) to 10-15% in many other fields. But doctors tend to believe their own error rate is lower - result of overconfidence and complacency. This persists in part because they are often not informed of their own errors.

Clancy - 10 Years After to Err is Human

Recognizing that improvements haven't been as successful as predicted. Suggestions: create a culture of safety, encouraging effective team communication, implement protected voluntary reporting to Patient Safety Organizations, require mandatory reporting of serious events, have AHRQ audit results, limit resident hours

Taylor - a steady-hand robotic system for microsurgical augmentation

Report on the development of a steady hand robotic tool designed to extend human's ability to perform small scale (sub-milimeter) manipulations requiring human judgment, sensory integration, and hand-eye coordination. Both surgeon and robot hold the tool, robot moves tool based on forces exerted by surgeon on tool and by tool on environment and uses this info to provide smooth, tremor free precise positional control and force scaling.

Sonderegger-Iseli - Diagnostic errors in three medical eras: a necropsy study

Researchers analyzed major and minor discrepancies between clinical diagnoses and necropsy findings to assess changes in diagnostic accuracy over 20 years. Found that the frequency of major diagnostic errors was halved over 20 years, likely because of improved clinical skills and new diagnostic procedures. Driven primarily by improved performance in cardiovascular diagnosis. Issues with study: classification of errors subject to interpretation (20% were reclassified); using autopsies means not capturing minimal errors

Give an example of the devices Vicky human factored.

Respiratory support pack - updated cue card in lid pocket since old one was hard to read; involved astronaut participation and iterative testing; redesign shaved 3 minutes off original 7 of completing emergency respiratory procedure Medical packs - studied unpacking and repacking equipment out of advanced life support and ambulatory medical packs for specific procedures; implemented changes to procedures, added labels, rearranged equipment; reported findings and new guidelines for med pack redesigns Digital stethoscope recordings - tested recordings to see if they provided clinically useful recordings despite ISS noise; compared with ground simulation recordings and found that sound quality differed depending on body sound, but with proper training, stethoscope could provide sufficient clarity.

Leape - IOM Figures Are Not Exaggerated

Response to the controversy and skepticism surrounding the figure. Why 98,000 figure might be understated: 1. relied on reported errors found in medical records (thus missed ones that were not recorded or were discovered after patient discharge, but unlikely that they found events that didn't exist), 2. only counted patients who died as a direct result of an error (left out patients who had other risk factors or ones where an error contributed to but did not cause death), 3. did not examine injuries that occur outside the hospital setting such as in ambulatory care units, surgical hospitals, or at home, and 4. these kinds of record reviews have been shown to give low estimates ***Recent research puts the number of accidental deaths closer to 400,000 and $24B cost to health care system from errors

Diagnostic Lab-in-a-Backpack

Rice Global Health students created an efficient and cost effective way to deliver quality healthcare to remote areas in developing countries. The portable lab in a backpack carries items such as a microscope for a physician to perform medical exams and can be tailored to fit specific regions. It also contains a power source for providing its own electricity and includes a solar panel for recharging the battery.

Weissman - Comparative Effectiveness Research on Robotic Surgery

Robot-assisted surgery is very expensive and is rapidly gaining in popularity of use - is it actually safer/better? For hysterectomies, added about $2000 extra to cost but wasn't clinically more effective. In some cases, it may have a shorter learning curve. But, could just spend resources on training physicians in laparoscopic techniques instead of buying robots. Another issue is marketing of robotic surgery - both by manufacturers and hospitals that have the robots - a bit questionable since there isn't a proven advantage.

Gawron - Medical Errors and HF Engineering: Where are we now?

Socially induced errors (ex. asking physician for a drug you've seen advertised) Active errors vs. latent errors (effect seen immediately vs. only evident later) Error taxonomy has like 40 different categories - very hard to decide which one is right for an error Preventability table issues - home error numbers look way too low (no reporting system); low in other areas with no mandatory reporting mechanism, like dr.'s office; operating room has highest because it has the most witnesses and clearest reporting

Barber - Should we consider non-compliance a medical error?

Sometimes intentional (violation, knowledge or rule based mistake), sometimes unintentional (slip or lapse). Theories of human error provide explanatory model of non-compliance. However, it is not actually a medical error

Gundry - Comparison of naive 6th grade children with trained professionals in the use of an automated external defibrillator

Study compared 1. time of entry into cardiac arrest scene to delivery of first shock and 2. pad placement using AEDs in trained professionals and untrained 6th graders. Electrode placement was adequate for all, and the speed of the children was only moderately slower than that of the professionals, suggesting that widespread use of AEDs requires only modest training.

Mykityshyn - Learning to use a home medical device: mediating age-related differences with training

Study comparing differential benefits of instructional materials for a blood glucose meter for younger and older adults. Performance was measured immediately and after a 2-week retention interval. Older adults trained using text-based manual had worst performance, but ones who received video training performed as well as younger adults. Type of training did not affect the performance of younger adults.

Svistun - vision enhancement system for detection of oral cavity neoplasia based on autoflorescence

Study found that autofluorescence enhances contrast between normal and neoplastic oral mucosa in fresh tissue resections, leading to better detection of the bad lesions.

Arnedt - neurobehavioral performance of residents after heavy night call vs. after alcohol ingestion

Study found that post-call impairment on attention, vigilance, and a driving simulation during a heavy call rotation is comparable to impairment associated with BAC of .04-.05%. Residents ability to judge this impairment may be limited and/or task specific. Compared with light call, residents were slower, made more errors, less able to maintain consistent speed and lane positions, and more likely to crash. There were no differences based on year of resident - you don't get "used to" sleep deprivation

Kortum - the impact of inaccurate internet health information in a secondary school learning environment

Study had science magnet high school students search the terms "vaccine safety" and "vaccine danger" using google and then answer questions regarding accuracy of the health info on returned sites. Despite over half of the links viewed being inaccurate on the whole, 59% of students thought that the sites were accurate on the whole. A high percentage of participants left with significant misconceptions about vaccines. Because of this, they were brought back in for a fact based video, after which 94% reported accurate information. Conclusions: allowing students to use the internet to gain info about medical topics should be approached with caution. Accurate info may need to be presented with anecdotal component to counter strong emotional message of internet zealots.

Wetterneck - technology characteristics predicting end user acceptance of smart IV infusion pumps

Study of nurse user perceptions and acceptance of smart IV pumps one month after implementation. 42% responded positively towards accepting the pump/ for 21 of 23 usability characteristics and performance options, more likely to be neutral than positive or negative. Highest positive perceptions were for ease of learning to operate pump and reliability. 6 characteristics predicted acceptance: perceptions that IV pump enhanced job effectiveness, made job easier, increased safety of care, and functioned as expected; perceptions that alarm messages were frustrating and interface was rigid were negatively correlated. Takeaway: highlighting improved patient safety may improve user acceptance of Smart IV pumps.

Drews - Development and evaluation of a just-in-time support system

Study simulated a medical emergency and compared subject effectiveness when using a paper-based NASA treatment protocol vs. a JITS system, which visualized protocol through simple animated graphics. Participants made more errors and took longer to stabilize the "patient" with the paper based instructions, demonstrating the benefit of a JITS system to instruct novices in unfamiliar tasks and improve survival rate of victims of cardiac arrest.

Zhang - A Cognitive Taxonomy of Medical Errors

Taxonomy can tie medical errors to underlying cognitive mechanism, explain how and why specific error occurs, and generate intervention strategies for each type of error Slips (result from automatic behavior) vs. mistakes (result of conscious deliberation) Taxonomy of slips and mistakes (separate trees): 1. Execution - occur at stages of goal, intention, action specification, and action execution 2. Evaluation - perception, interpretation, action evaluation *goal is general, intention is specific to device

Lamsdale - a usability evaluation of an infusion pump by nurses using a patient simulator

Using a hifi patient simulator, nurses evaluated advanced features of an IV pump in several scenarios after a training session with the pump. Nurses were required to use a think out loud protocol. The most common problems were with the "change mode" and "select new patient" features. Use of the on/off switch was a common strategy to clear pump info and escape incorrect paths. Dangers to patient ranged from non-hazardous to very hazardous. A number of design recommendations were made to address problems with use.

Barger - Impact of extended-duration shifts on medical errors, adverse events, and attentional failures

Web-based survey of interns looking at association between # of extended duration shifts worked in a month and medical errors, adverse events, and attentional failures. Commission of an error was 3 times more likely if dr. worked 1-4 extended shifts in a month and 7 times more likely if 5 or more extended shifts in a month. Fatigue related adverse events were 7-8 times more likely if doctor worked extended shifts, and fatigue related events contributing to death of patient increased 300%. Interns were more likely to fall asleep during lecture, surgery, while talking to or treating patients, and during rounds. Issues with study: participants self-selected (might attract people who are bothered by sleep deprivation); only analyzed those who completed all surveys (makes the above problem worse as only those most interested/motivated will do that); questions used to elicit error responses may have lead the witness; hindsight bias of easier to blame fatigue than carelessness; all errors self-reported with no independent verification; adverse events that almost happened might not be captured and other errors might not be recognized