BMI 501 Final Flashcards

pg 4 of: Chapter 5. Understanding Adverse Events: A Human Factors Framework

“Human Error—A Troublesome Term

While one frequently finds references to human error in the mass media, the term has actually fallen into disfavor among many patient safety researchers. The reasons are fairly straightforward. The term lacks explanatory power by not explaining anything other than a human was involved in the mishap. Too often the term ‘human error’ connotes blame and a search for the guilty culprits, suggesting some sort of human deficiency or lack of attentiveness. When human error is viewed as a cause rather than a consequence, it serves as a cloak for our ignorance. By serving as an end point rather than a starting point, it retards further understanding. It is essential to recognize that errors or preventable adverse events are simply the symptoms or indicators that there are defects elsewhere in the system and not the defects themselves. In other words, the error is just the tip of the iceberg; it's what lies underneath that we need to worry about. When serious investigations of preventable adverse events are undertaken, the error serves as simply the starting point for a more careful examination of the contributing system defects that led to the error. However, a very common but misdirected response to managing error is to “put out the fire,” identify the individuals involved, determine their culpability, schedule them for retraining or disciplinary action, introduce new procedures or retrofixes, and issue proclamations for greater vigilance. An approach aimed at the individual is the equivalent of swatting individual mosquitoes rather than draining the swamp to address the source of the problem. “

Human-system interfaces include medical devices, controls, displays, equipment location, software controls, design and usability. Can increase medical error if design and usability are not ideal, particularly in emergency situations. Must take into account capability and knowledge of users, including users in a home setting. Implementation of new systems is not only an IT project, the users must master use.

Can decrease medical errors with software controls and easy to understand displays.

External factors - economic pressures such as increase in outpatient care, versus inpatient care. Science and technology advances that make devices smaller, easy to use. Changing demographics with consumers with more computer literacy.

External representations: quickly determine the state of the device

Functional components of an EHR (p.394)

• Integrated view of patient data includes records of patient visits, tests, labs, imaging, patient history, can also include dashboard summaries.
• Clinician order entry for provider to submit orders, often by order sets
• Clinical decision support such as alerts, reminders, calculators.
• Access to knowledge resources such as info buttons, calculators
• Integrated communication and reporting support can include transfer of information to another clinician or an HIE

Options for recording physician entered data

Three current mechanisms (p.406)

• Transcription of dictated or written notes
• Clinic staff transfer or coding of some or all of the data from a paper form
• direct entry by physicians in the EHR - both structured fields and free text

How to achieve cutting edge solutions? (pp. 417-9)

• Provide better workflow and anticipation of users' information needs (achieve this through observational studies)
• Increase system usability
• Provide alternate / automated data entry techniques
• Patient-entered data (history, symptoms, outcomes)
• Patient-generated data (uploaded from PHR)
• Data provided directly from equipment & monitoring devices

Primary purposes of CDS systems (pp.644-5):

• Aid in diagnosis ( problem solving )
• Aid in management / choice of treatment ( decision making )

Difference between evoking strengths and frequency weights

• ??? [didn't find this in the book or my notes - maybe in the videos?] Patricia sent email to Dr. Kaufman for clarification. Here is what he said

“I may have to remove that question or more likely, that subquestion. Internist was an important topic and system in the history of CDS, but seems to have been removed from the updated text. What a shame.”

http://www.fpnotebook.com/mobile/Manage/Computer/ClnclDcsnSprt.htm

“Internist-I (1974) followed by Quick Medical Reference (1980)

1. http://en.wikipedia.org/wiki/Internist-I
2. Scoring and heuristics System
3. Disease profiles contained findings (history, exam, lab) and their evoking strength (0-5) and frequency (0-5)
1. Evoking strength is equivalent to Positive Predictive Value (PPV), the likelihood that a disease is present if a finding is present
2. Frequency of a finding is similar to Specificity; an absent finding that has a high frequency, would rule out the condition”

CDS in public health contexts:

• Review of surveillance data could lead to diagnosis of community health problems (p.505)
• Public databases of immunization records could be used to identify individuals due for vaccinations. This info could feed into EHR systems. (p.511)

Dr. Murcko’s presentation CDS purposed:

• monitor (alerts, reminders)
• organize information
• find information
• help make decisions
• perform calculations

Researchers study detectability using real surveillance data collected during real disease outbreaks. RODS can be used to gather continuous data collection, organize the data for analysis, perform data analysis, outbreak or case detection, make a decision, and implement a response action.

Detection systems looking for abnormalities - ED complaints (easy to attain, early in disease process, low diagnostic precision), OTC drug sales (available early in the disease process, and electronically available on a daily basis), school absenteeism, internet searches, provider hotline volumes, lab results.

Prior studies have taken note of correlations between thermometer sales during influenza outbreaks along with pediatric electrolytes for pediatric respiratory and diarrheal outbreaks. Another example is seen by Stirling et al in which a large waterborne outbreak of a parasite (Cryptosporidium) led to a peak in the sales of OTC diarrhea remedies weeks before precautionary drinking water advisories were released.

Mining of social media can foster a faster detection of an outbreak, wouldn’t have to wait for ED complaint reports.

Detection Algorithms Include:

Univariate, Multivariate, Spatial Scanning, Specialized Algorithms

Multivariate algorithms search for anomalies using multiple data streams. An example of this can be seen through GI emergency department visits and the sale of OTC antidiarrheal products. Challenges include false alerts and the computation time.

Tier 1 is the health status that is achieved which varies depending on illness or injury. The first level with in this tier is survival/mortality rate. The second level is the degree of health recovery. The second Tier revolves around the recovery process. The first tier is the time it takes to obtain best recover outcome, while the second tier addressed the short-comings of the the quality of care. This second level may involve a misdiagnosis, anxiety, discomfort, failed treatment, etc. The third tier involves the sustainability of health. The first level addresses recurrence of the original disease or associated longer term complications. Measures of time to recurrence and the seriousness of recurrence would fall here. The second level addressed additional complications or illness that were due to the original treatment, hospital visit, surgery, etc. Staff infection pandemic for example. [http://www.nejm.org/doi/suppl/10.1056/NEJMp1011024/suppl_file/nejmp1011024_appendix2.pdf]

A few of the elements in outcome measurements efforts should should be to include at least one outcome dimension at each tier of the hierarchy, and ideally one at each level. Measures should be important to the patient, variability, frequency, and practicality. [http://www.nejm.org/doi/suppl/10.1056/NEJMp1011024/suppl_file/nejmp1011024_appendix2.pdf]

Elements of preferred outcome measure - value (cost compared to outcomes), efficiency, quality care process, health circumstances (patient’s initial condition), short and long term health.

• Health status Achieved - survival/mortality rate
• The first level, survival, is of overriding importance to most patients and can be measured over various periods appropriate to the medical condition; for cancer, 1-year and 5-year survival are common metrics. Maximizing the duration of survival may not be the most important outcome, however, especially for older patients who may weight other outcomes more heavily
• The second level in Tier 1 is the degree of health or recovery achieved, which normally includes dimensions such as freedom from disease and relevant aspects of functional status.

• Recovery process - addresses shortcoming of quality of care
• The first level is the time required to achieve recovery and return to normal or best attainable function, which can be divided into the time needed to complete various phases of care. Cycle time is a critical outcome for patients — not a secondary process measure, as some believe. Delays in diagnosis or formulation of treatment plans can cause unnecessary anxiety. Reducing the cycle time (e.g., time to reperfusion after myocardial infarction) can improve functionality and reduce complications.
• The second level in Tier 2 is the disutility of the care or treatment process in terms of discomfort, retreatment, short-term complications, and errors and their consequences.

• Sustainability of health -
• The first level is recurrences of the original disease or longer-term complications.
• The second level captures new health problems created as a consequence of treatment. When recurrences or new illnesses occur, all outcomes must be remeasured.

Shortened:

health status: 1. survival 2. degree retained

recovery: 1. time required 2. "disunity of care"

sustainability: 1. recurrence 2. new problems

Methods and Theories from HCI (pp. 111-2)

• seed basic research findings that can illuminate dimensions of design (e.g., attention and memory, aspects of the visual system) --> visual imaging systems
• provide an explanatory vocabulary for characterizing how individuals process and communicate health information --> medical knowledge bases, digital libraries, health literacy
• present an analytic framework for identifying problems and modeling certain kinds of user interactions --> evidence based clinical guidelines
• characterize the relationship between health information technology, human factors and patient safety --> alerts, reminders
• provide rich descriptive accounts of clinicians employing technologies in the context of work -->
• furnish a generative approach for novel designs and productive applied research programs --> decision support, visual/graphical representations (p.147)

Steps to promote adoption

• For the product team:
• Ensure the product meets user needs by starting with observational studies
• Perform usability testing to ensure efficiency of use & low error rate
• For the medical team:
• Leadership must support and evangelize adoption
• Provide information sessions and training

Consequences of failure to take workflows into account

• lack of adoption by clinicians > failed implementation > financial loss
• decreased patient safety
• increase time spent in system by clinicians > frustration & lack of productivity

From Dr. Kaufman’s presentation

• Cognitive theory
• external representations and visualizations
• usability evaluation methods (interview, cognitive walkthrough, heuristic evaluation, usability testing

Steps can include requirements analysis, usability testing, interviewing users, understand the environment. Design process should be iterative. Failure to do so can lead to unintended consequences, including problems with adoption, productivity and patient safety.

“The Institute of Medicine’s (IOM) National Roundtable on Health Care Quality documents three types of quality problems—overuse, underuse, and misuse”

“Four key aspects of the current context for health care delivery help explain the quality problems outlined above: the growing complexity of science and technology, the increase in chronic conditions, a poorly organized delivery system, and constraints on exploiting the revolution in information technology.”

“Six specific aims for improvement. Health care should be:

• Safe—avoiding injuries to patients from the care that is intended to help them.
• Effective—providing services based on scientific knowledge to all who could benefit and refraining from providing services to those not likely to benefit (avoiding underuse and overuse).
• Patient-centered—providing care that is respectful of and responsive to individual patient preferences, needs, and values and ensuring that patient values guide all clinical decisions.
• Timely—reducing waits and sometimes harmful delays for both those who receive and those who give care.
• Efficient—avoiding waste, in particular waste of equipment, supplies, ideas, and energy.
• Equitable—providing care that does not vary in quality because of personal characteristics such as gender, ethnicity, geographic location, and socioeconomic status.”

Source: (You’ll need to make a free account on the site to download, Chapters 1 and 2) http://iom.edu/Reports/2001/Crossing-the-Quality-Chasm-A-New-Health-System-for-the-21st-Century.aspx --------

or Dr. Kaufman’s Systems Engineering presentation

Primary causes:

rapid advances (increasing complexity), increase in specialization (disconnected silos), patient population needing chronic care. Also structure of US market for health care services, underinvestment in information and communications technology, inability or unwillingness to take advantage of engineering based system design analysis and management tools.

Dimensions of quality:

Safe, effective, patient-centered, timely, efficient, equitable

Measures we can take:

Measures to improve quality - use concurrent engineering to overcome silos, use teams of specialists. Understand stakeholder needs, streamline processes. Systems control

Radiologic Process Tasks:

1. Evaluation by clinician of clinical problem and determine need for imaging procedure
2. Procedure is requested and scheduled w/ indication
3. Imaging Procedure carried out
4. *Radiologist review image in context of clinical history (The radiologist reviews the images in terms of the clinical history and questions to be
   answered and may manipulate the images. This task actually involves two interrelated subtasks: (a) perception of the relevant findings and (b) interpretation of those findings in terms of clinical meaning and significance)
5. *Radiologist creates report (The radiologist creates a report and may otherwise also directly communicate the results to the referring clinician, as well as making suggestions for further evaluation as needed. The process may then be repeated if a subsequent procedure is considered
   to be helpful).
6. Quality control carried out - look at waiting times, workloads, etc
7. *Continuing education and training for the physician (Continuing education and training are carried out through a variety of methods,
   including access to atlases, review materials, teaching-file cases, and feedback of subsequently confirmed diagnoses to interpreting radiologists)

How can these tasks be enhanced through IT?

4<==Technology largely impacts the manner in which groups operate and think. When using an effective medical information system, physicians should be able to gather information more systematically and efficiently. Information technologies may alleviate some of the cognitive load associated with a given task and permit them to focus on higher-order thinking skills, such as diagnostic hypothesis generation and evaluation.

5<==Information technology can significantly improve the manner in which radiologists create patient reports. Within the healthcare field, comprehensive, longitudinal, clinical databases (i.e. a paperless, complete medical record) significantly enhance the integration of healthcare delivery. Will provide a complete picture of the patient's medical history and will assist in avoiding duplicate tests being performed and unfavorable drug interactions.

7<== In an attempt to continue physician education and training, computer-assisted learning systems can be utilized in order to to provide physicians with quick access to references and new content. Most importantly, these lessons are highly individualized and may exercise a physician's knowledge and decision-making capabilities within a nonthreatening environment.

• Image enhancement (sharpen, highlight edges, improve contrast)
• Image restoration (remove distortions and blurs, despeckle, smoothing, remove artifacts)
• Image segmentation (find edges for measuring, remove details)

Needs to be complementary

• Computer Aided Detection (CAD) assists physicians by placing markers in images where physicians are more likely to find useful information. This can reduce cognitive load and review time by helping the physician to perform a more targeted review.

Bioinformatics create new diagnostic and prognostic information sources that influence clinical care. Single nucleotide polymorphisms (SNPs) and other genetic markers from a patient will be highly specific and sensitive indicator of the subtype of disease and of that subtype’s probable responsiveness to different therapeutic agents. They also use as prognostic tools to guide health management based on individualized risk of disease in the future. Historically, guidelines for a healthy lifestyle have been presented as universal. Clinical care guideline have been based primarily on macroscopic symptoms and qualities that can be observed in the course of a physical exam or reported by the patient.

(BMI book ch 24 Bioinformatic p.697-698, ch 25 translational bioinformatics p.724, 726)

Due to large genetic sequence information, the EHRs need to adopt new standards for describing clinically significant genomic information as well as, to develop the way to;

1. Maintain separation of primary molecular observations from the clinical interpretations of those data
2. support human-viewable formats and machine-readable formats in order to facilitate implementation of decision support rules.
3. Maintain linkage of molecular observations to the laboratory methods used to generate them

(any other suggestion?)

(http://www.ncbi.nlm.nih.gov/pmc/articles/PMC3328607/)

GINA - Genetic Information Nondiscrimination Act of 2008

HIPAA protects an individual identifiable health information held by covered entities and give patients right in respect to that information. However the privacy rule also allows researchers access to this information after it has been de-identified. The same thing will probably happen with GINA which is to assure that health coverage and employment will not be allowed to discriminate based on genomic information.

To build trust with its patients and meet HIPAA requirements, a healthcare organization (HCO) should adopt a three-pronged approach to securing information. First, HCO need to designate a security officer and develop uniform security and confidentiality policies, including specification of sanctions and to enforce these policies rigorously. Second the HCO needs to train employees so they understand the appropriate uses of patient-identifiable information and the consequences of violations. Third, the HCO must use electronic tools such intrusion detection, access control and audit trails not only to discourage misuse of information but also to inform employees and patients that people who access confidential information without proper authorization or a “need to know” can be tracked and will be held accountable.

(Ref BMI book ch 14 p. 454-455)

“Immunization registries are public health tools that facilitate high immunization coverage rates and minimize children being over immunized.” Contain record of children from specific geographic area, info about what immunizations they have had. Usually entered in to record at birth or first immunization(first immunization is for Hep.B at birth).

Registry model could be used to also classify those with diseases and track prevalence and spread of disease. Could contain registry of people with AIDS or common chronic spreadable disease. Registry could also be applied to people’s genomic information. Could be used to track people’s health patterns - how often they see the doctor, what their diet is like, how often they exercise, how many hours a week they work ect. This info could be quantified and mined to produce a relationship of patient’s habits to their health.

Other thoughts??

Registry model has ethical limitations and considerations, like if patients want their information in a format like that. Who would have access to it and for what reasons ect.

(Textbook p:697)

The sequence of a gene involved in disease may provide the critical information that we need to select appropriate treatments. For example, the set of genes that produces essential hypertension may be understood at a level sufficient to allow us to target antihypertensive medications based on the precise configuration of these genes. Clinical trials now use information about genetic sequence to define precisely the population of patients who would benefit from a new therapeutic agent. Finally, clinicians may learn the sequences of infectious agents (such as of the Escherichia coli strain that causes recurrent urinary tract infections) and store them in a patient’s record to record the precise pathogenicity and drug susceptibility observed during an episode of illness. Raw sequence information, whether from the patient or the pathogen, is meaningless without context and thus is not well suited to a printed medical record Like images (CAT scans, MRIs), sequence data can come in high information density and must be presented to the clinician in novel ways. As there are for laboratory tests, there may be a set of nondisease (or normal) values to use as comparisons, and there may be diffi culties in interpreting abnormal values. Fortunately, most of the human genome is shared and identical among individuals; less than 1 % of the genome seems to be unique to individuals. Nonetheless, the effects of sequence information on clinical databases will be significant. As an example, Kaiser Permanente and UCSF recently completed identifying genome wide genetic variants in 100,000 individuals linked to their electronic health records. 1 The EMERGE research network is specifically focused on genetic discovery and replication using electronic medical records and genetics.

Plan for usability evaluation: start with quick low-cost methods, increase in rigor as necessary (see pp.128-133)

• [Inspection] Start with a heuristic evaluation including input from several reviewers.
  This is a quick way to find obvious usability issues, which can be communicated to the software development team for repairs.
• [Observation] Identify common tasks performed with the EHR system in the ICU, and perform a formal usability test, observing representative users.
• Usability testing commonly employs video capture of users performing the tasks as well as video-analytic methods of analysis. It involves in-depth testing of a small number of subjects ( Page: 133)

Other factors to consider: maybe the problem is not really the software. The problems could have to do with the physical environment, work processes, lack of training, staff resistance, location and number of EHR workstations, etc.

• [Observation] If findings from the first two tests do not provide enough answers, carry out a full-fledged contextual observation of work processes, to see if other inefficiencies become obvious.

Info:

There are numerous benefits to real-time data which includes a larger information platform in which large quantity of data may be garnered in real-time. Such sites include popular blogspots along with Twitter or Facebook. In considering adverse Drug Reactions(ADR)

Public Health info like spreading of diseases or disease prevalence

Technique:

Text mining - Machine Learning==> machine learns a model by generalinzing from training data

Process of discovering and extracting new knowledge from unstructured data

Mean: Wearable computers creates new opportunities for local environment data collection, personal vital sign and activities and more through disease monitoring.

Advantages: We can expect the use of these massive data sets from wearable devices to impact how medical care is personalized and patient decision making.

Disadvantages: However they posing threats to personal privacy and reliable health data.

Challenges: Interoperability, security, quality, health literacy, availability (digital divide)

Other factors: socio-economic, genetics, behavior

Reasons for variability:

Why some regions do well and others poorly:

• Access to healthcare (not everyone has insurance)
• Time to treatment (availability of staff and services)
• Not all treatment facilities can afford the latest technology (depends on local economy)

Explain the differences:

• Pay for service model leads to variable pricing strategies
  • different incentives
  • different affordability
  • Variable processes / practices might lead to different outcomes
• • Some facilities are connected to others, some are not (data sharing issues)

How big data can help:

• Find the high value performing facilities and discover what they do differently. Systematic view of healthcare - silos of knowledge domain, care domain, payer domain. Not connected and only some information flows through, few feedback loops. No integration, no coordination. Low value (quality/cost over time). Domains should overlap, with the patient in the center.

Value (value=quality/cost=(outcomes, safety, service)/(cost over time))

Individualized medicine (for the individual, and targeted populations)

Science of health care delivery (systematic coordination for complex conditions)

Integration and coordination (of people & data)

He also mentions: insurance for all (equal access to care), and pay for value

Cortese lecture, slide 4: Components of a Learning Healthcare System:

• Shared Vision
• Professionalism
• Information Technology
• Systems Engineering

An advantage of increasing the number of physicians within the United States is that there will be a decrease in the amount of time that patients will have to wait for any appointments or elective surgeries. A disadvantage of such an action includes the fact that through an increase in the number of physicians, there will also be an increase in the number of fee-for-services which largely requires for patients to pay for services out of pocket. Along with such a disadvantage, there will also be a higher health cost per capita.

Human factors deals with safety in a system and how people work with the system. Human factors includes information about human behavior, abilities, limitations, and design of tools for human use (p. 133 in book). Human computer interaction deals specifically with design and usability (p. 128 in book).

Public health informatics focuses upon the health of a population, while clinical informatics is focused upon individual health In considering the strategy of public health, it is prevention while clinical informatics strategy is based upon treatment. Public health informatics may take place in any setting while clinincal informatics is an encounter. Lastly, public health informatics operates with the government while clinical informatics operates with both the government and private sectors.

It emphasizes sequential processes. It also assumes that the reviewer understands something of the typical user's existing knowledge and practices.

---The cognitive walkthrough (CW) is a cognitive task-analytic method that has been
applied to the study of usability and learnability of several distinct medical information
technologies (Kushniruk et al., 1996). The purpose of a CW is to characterize the cognitive processes of users performing a task. The method involves identifying sequences of actions and goals needed to accomplish a given task. The specific aims of the procedure are to determine whether the user’s background knowledge and the cues generated by the interface are likely to be sufficient to produce the correct goal–action sequence required to perform a task. The method is intended to identify potential usability problems that may impede the successful completion of a task or introduce complexity in a way that may frustrate users.

Shared decision-making (SDM) is an approach where clinicians and patients communicate together using the best available evidence when faced with the task of making decisions, where patients are supported to deliberate about the possible attributes and consequences of options, to arrive at informed preferences in making a determination about the best action and which respects patient autonomy, where this is desired, ethical and legal.(Wikipedia)

It is important when treating a cancer patient because the patient may value living the rest of their life without treatment, rather than being treated - living a little longer - but having their quality of life be worse during those years they are alive/being treated.

Kaufman’s presentation on Systems Engineering - breaking down silos, distributed decision making.

Book chapter 4 - limitations of each decision making, cognitive load, different levels of medical cognition, different levels of expertise min medicine and the reasoning process, distributed cognition.

Kaufman’s Consumer Health presentation - clinical decisions involve judgment, interventions have risks and benefits, no single right answer for everyone, health care providers cannot guess at what patients value, health care providers cannot assume what is in patient’s best interest, evidence-based practice movement.

Cost is a major pressure that is being put on consumers and patients, and the more the actively participate in maintaining their health the less they will have to pay out in the future.

Modern technological advances give them the obligation to become more involved through education, biosensor, medical portals etc. it says factors plural it is s stretch though. thoughts?

Partitioning of patient care and coordination of medical records?

(p. 135 in book) Active failures are immediately felt, are at the “sharp end” of the process. Latent conditions are less visible, may not be evident for some time, and weaken the system’s defenses, make errors possible.

According to Dr. Murcko’s presentation:

alerts, reminders

clinical guidelines

order sets and checklists

reports and summaries

dashboards

templates

pertinent references at key points in the workflow - info buttons

Clarified this with Dr. Kaufman, he says this question is “What are three purposes or uses for clinical decision support.”

from Dr. Murcko’s presentation:

find information, help make decisions, perform a computation, manage and optimize process and workflow, monitor (alerts, reminders, feedback), organize information, focus attention.

According to Dr. Murcko’s presentation:

personal/professional - need perceived value of CDS

organizations - need workflow, smoothness of integration, leadership

business - incentives for users, ROI

technical - software, hardware, interoperability

According to Dr.Kaufman`s presentation: (Human Factors and HCI):

• Ability to ignore extraneous information and focus on relevant inputs.
• Performance typically declines as the number of sources of information increases
• Humans can only process information at a finite rate
• Information overload contributes to errors

A dashboard is a graphical summary, and can be used in an EMR to provide a summary of a patient’s health and healthcare alerts. Dashboards can also be targeted to show information relevant to the clinician's concerns; a neurologist will be looking for different information than an endocrinologist.

Contrast - difference in image intensity, bits per pixel

Spatial - ability to distinguish, sharpness, pixels per area

Temporal - time needed to create, frames per second

--A form of mental representation that enables one to understand how something in the world works. One can run a mental model to predict future states of a system (what happens when I click this link) or to explain the cause of a change in the state of a system (why did my computer crash)

A mental model is an explanation of someone's thought process about how something works in the real world.

(p. 119 in book) Mental model refers to how individuals form internal models of systems. They are always incomplete, imperfect, and subject to the processing limitations of the cognitive system. They are useful in understanding HCI.

Ontologies are used in the process of organizing biosurveillance data to set up infrastructures for data analysis. The data comes from various source such as laboratories, hospitals, pharmacy etc. Ontologies set up concepts for different data types and define the relationships between those concepts in order to effectively utilize and retrieve data.

(BMI 501 Biosurveillance slide page 3 + opinion)

Lack of standards for interoperability (Greenes lecture)

Lack of common incentives / conflicting goals (Porter paper)

Interoperability

patient privacy

ethical concerns (who has access, when, and why)

big data (storage/retrieval negotitive concerns)

Concurrent engineering is an approach in which a particular problem may be cut into separate fragments to be worked on by various groups and ultimately the problem will be solved through the collective development of an effective solution. It can be used to overcome silos of function and responsibility and additional benefits include: a decrease in the number of design changes, a reduction in the design-to-delivery time, a decrease in defects, lower costs, improved teamwork, cross-training, improved quantitative and qualitative characterization of processes and systems.

Systematic technique for failure analysis. developed in 1950s to study problems with military systems. First step in system reliability study. Reviews as many components, assemblies and subsystems as possible to identify failure modes and their causes and effects.

Failure modes: ways, or modes, that something might fail(any error/defects)

Effect Analysis: refers to study of consequences of failures

Definition of failure-mode effects analysis - systematic, proactive method to evaluate processes to identify where and how it might fail. Asses the impact of these failures

Source : http://www.ihi.org/resources/Pages/Tools/FailureModesandEffectsAnalysisTool.aspx

To accomplish this goal, people need to develop data mining process by:

-Must first define what readmission is (i.e. it is within x amount of days)

-Understanding data which is understand what data is in EHR.

-Preparation for data which is identify what risks for readmission are such as age, comorbidity, critical laboratory results.

-Modeling algorithm which is use of machine learning to develop a predictive model.

-Evaluation of algorithm which is evaluate performance of the learn model in test data and redevelop the learn model.

- Cooperation to obtain “loss to follow-up:” when the patient is readmitted, but somewhere else and we don't know about it.

-use past cases and data on readmission to infer about the future

• information retrieval (gather relevant text), document or paragraph level
• Text processing (lexical and grammatical level), tokenization, stemming
• information extraction (identify and extract specific types of entities or information, and the associations among them)

Pragmatic: subfield of linguistics and semiotics studies ways in which context contributes to meaning

Semantic Analysis: process of relating syntactic structures to the level of writing as a whole. Also involves removing features specific to particular linguistics and cultural contexts.

-So: Pragmatics = impact of context and of intent of the speaker on meaning

Semantics = meaning of words, phrases, and sentences

(p. 257 in book) Abbreviated words, misspelled words, typographical errors, word sense disambiguation, semantics,

Science that seeks to explain how variations in the human genome can affect the individual response to drugs.

Due to variability in patient response to this drug, including genetic variations, the dose is based on an algorithm that includes the gene encoding for a specific enzyme that metabolizes the drug and a gene encoding for the target receptor of the drug.

SNP is the DNA sequence variation that occurs when a single nucleotide in the genome sequence is altered. In other words, it is a mutation. An SNP can cause the transcribed amino acid to differ and intern the translated protein to slightly differ, possibly changing structure and function.

--A DNA sequence variation, occurring when a single nucleotide in the genome is altered. For example, an SNP might change the nucleotide sequence AAGCCTA to AAGCTTA. A variation must occur in at least 1% of the population to be considered an SNP.

Noise in samples which needs lots of steps in preprocessing samples and noise in sequencing data which needs several analytical steps to quantify differentially expressed genes

High error rates (Sequencing errors and alignment errors) for long reads of base pair (3-7000 bp)

(BMI 501 Bioinformatic lecture slide p.37, 44-46 + opinion)

From Peng’s presentation slide 26/27 - data deluge, computational limitations, incomplete sequences, cost

Problems are the lacks of ability to interchange or transfer images and associated information between devices manufactured by different vendors that display a variety of digital image format.

(BMI book chapter 7 page 238-239)

Prior to Digital Imaging and Communications in Medicine (DICOM), the public was in need of a standard method that would enable them to transfer their images and associated information between devices that featured a broad range of digital imaging formats. Such a high level of variety resulted from the fact that these products were being manufactured by various vendors. DICOM provides a specified image-related management information exchange, with the potential to interface to hospital and radiology information systems.

- Spatial Resolution (Sharpness of the image); A measure of the ability to distinguish between two close objects on an image, related to the number of pixels per image area

– Contrast Resolution; A measure of the ability to distinguish small differences in image intensity. Generally, related to the number of bits per pixel

– Temporal Resolution; A measure of the time needed to create an image. Generally, related to the number of frames per second

(BMI 501 imaging informatics slide page 40-45) lectured on 11/25/14

• Clinical trials: finding qualified candidates (p.415)
• Quality Reporting, both for internal improvement and for external reporting (pp.415-6)
• Retrospective research (p.416)
• Administration: understanding practice patterns and resource utilization (pp. 416-7)

1.Outcome measures in different contexts, such as quality improvement, public reporting, and incentive programs, can be different.

1. A variety of health care services can be major determinants to an outcome. eg. for heart attack outcomes, these could include services delivered by emergency medical technicians in the field, emergency room (ER) teams, inpatient and cardiac catheterization laboratory staff

3.The need to account for all factors that influence a patient's health outcome such as comorbidities, behaviors and other important technical questions such as the sample sizes.

(This is the best source I can find. Not sure it is a good answer.)

(http://www.qualitymeasures.ahrq.gov/tutorial/HealthOutcomeMeasure.aspx)

Porter article - current measures dead @ discharge, readmission w/in 30 days. Not included mortality after 30 days, sustainable health, functional capacity, health related costs, problems with prescription drugs.

Advantages and disadvantages of Big Data

+ Big data can be used to support knowledge discovery and associations of health factors such as finding biomarkers and prediction of disease outbreaks in biosurveillance in timely manner.

Disadvantage

- Unknown population representation

- Issues of data quality

- Typically not very multivariate (at the person level)

- Privacy and confidentiality issues

- Difficult to assess accuracy and uncertainty

(http://www.cdc.gov/nchs/data/bsc/bscpres_schenker_091913.pdf) + opinion

Same as question 30

Don’t follow patients, don’t share information.

-Summary of the grand challenges of clinical decision support

Improve the human-computer interface

Disseminate best practices in CDS design, development, and implementation

Summarize patient-level information

Prioritize and filter recommendations to the user

Create an architecture for sharing executable CDS modules and services

Combine recommendations for patients with co-morbidities

Prioritize CDS content development and implementation

Create internet-accessible clinical decision support repositories

Use freetext information to drive clinical decision support

Mine large clinical databases to create new CDS

Sittig DF1, Wright A, Osheroff JA, Middleton B, Teich JM, Ash JS, Campbell E, Bates DW. Grand challenges in clinical decision support. J Biomed Inform. 2008 Apr;41(2):387-92. Epub 2007 Sep 21.

http://www.ncbi.nlm.nih.gov/pubmed/18029232

(pp. 643-4)

1. use information about the current clinical context to bring back relevant info: infobuttons
2. Provide patient or situation-specific recommendations: alerts, reminders, order sets
3. Organize information for decision-making: dashboards, templates, reports

EMR - Electronic medical record, information about a patient’s health status and health care, also includes information management tools, clinical reminders, links to knowledge sources, analysis of aggregate data. Maintained by one health care organization.

EHR - Electronic health record, lifetime health status, multiple users, evidence based decision support, outcome reporting, quality management.(interoperable between health care organization)

PHR - Personal health record, patient oriented instead of provider oriented, benefits include empowerment, partnership, better decisions, more engaged patients.

(from book)

- Integrated view of patient data

- clinical order entry

- clinical decision support

- access to knowledge resources

- integrated communication and reporting support

(from notes)

- lab results

- notes

- discharge summary

-user choice; prefered mode of data entry by clinicians

-expressivity; narrative note to express patient history and other observations

-flexibility; captures clinical reasoning and idiosyncratic aspects of patient record

-more accurate (few errors due to drop down menu selection or checking boxes)

(BMI 501 EHR slides lecture on 9/18/14)

RNA

mRNA

Proteins

Pathways (from my notes)

Bile(found it in an article)

Costs

Leadership

Usability

Privacy and Security

Standards

Meeting users’ information needs

(additional points - From the slides)

1. perceived lack of return of investment
2. technical and logistical challenges
3. cost of conversion/ implementation

Advantages: genetic related disease diagnosis, individualized therapy

Disadvantages: discriminating people based on genetic data, selection of newborns (by trial and errors)

(my opinion)

Disadvantage: some people do not want to know their genetic predispositions; ignorance is bliss

The purpose is to standardized communicating and exchanging health information between independent computer applications (EHRs). They reduce cost of interfacing between different systems.

(p 12 and 220 in BMI book)

Researchers use text mining methods to build predictive models from social media posts. They use keywords to collect social media posts containing medication and adverse events words. Then they annotate terms to develop training data and use them to develop lexicon based, rule based or machine learning based algorithms that can classify and extract actual adverse drug reactions from posts.

(Ref: NPL in biomedical text mining lecture, slide page 41 -46)