Term
| IOM report/follow up: statistics |
|
Definition
| 44,000-98,000 poeple die in hospitals each year as a result of medical errors that could have been prevented. 17% adverse effect are diagnosis errors |
|
|
Term
| IOM report/ follow up: problems |
|
Definition
| drug events and improper transfusion, surgery injuries, wrong site surgeries, suicides, restraint-related injuries, death, falls, burns, pressure ulcers and mistaken patient identities. |
|
|
Term
| IOM report/ follow up: medical error costs |
|
Definition
| $17 billion-$29 billion per year in hospitals nationwide |
|
|
Term
| IOM report/follow up: Medical error contribution |
|
Definition
- from the nonsystem (decentralized and fragmented system
- patients see multiple providers in different settings, none of which have access to complete patient history/information
- little attention focused on prevention of medical errors by licensed and accredited health professionals.
- financial incentives are not aligned, therefore leading to an unsafe healthcare system |
|
|
Term
| IOM report/ follow up: Why did the 2013 study estimate the deaths to be so much higher than the IOM report? |
|
Definition
- IOM estimates 98,000 deaths based on a larger sampling of medical charts.
- The bar for identification of a PAE was higher than the four modern studies
- outdated (3 decades old)
- Did not include errors of omission (things that people didn't do that ended up killing someone)
- 2013 study estimates 210,000-400,000 deaths based on the GLOBAL TRIGGER TOOL (screening method used to find infection, injury, and error in medical reports) of which the number was averaged over four studies.
- Medicine is more complex now (practices and technology), leading to more mistakes.
- The 2013 study is more comprehensive and evidence-based.
- The GTT is better able to identify adverse events rather than just general reviews by physicians.
- The GTT makes it easier to find errors, getting better at detecting errors to identify what actually happened more accurately.
- This is a process issue, not a people issue. |
|
|
Term
| IOM report/follow up: How did organizations, especially government agencies, respond? |
|
Definition
- Public and private purchasers (businesses supplying insurance for their employees) must make safety a prime concern in their contract decision.
- Congress launched a series of hearings on patient safety and donated $50 million in 2000 to AHRQ in order to help reduce medical errors.
- National Academy for State and Health Policy approached the executive and legislative branches on ways to help enforce mandatory hospital error reporting.
- Private sector (leapfrog) unveiled a market-based strategy to improve quality and safety. |
|
|
Term
| PLAN-DO-STUDY-ACT: definition |
|
Definition
- A problem solving model to improve a process or carry out a change.
- A form of CONTINUOUS quality improvement (repeatedly improving a process to reach desired outcome.
- W. Edwards Deming: Father of quality evolution (Deming Cycle)
- Introduced to Deming by his mentor, Walter Shewart |
|
|
Term
|
Definition
| Plan: plan the test or observation, including a plan for collecting data. Objective, questions & predictions (why), plan to carry out the cycle (who, what, where, when), plan for data collection |
|
|
Term
|
Definition
| Do: Try out the test on a small scale. Carry out the plan, document problems and unexpected observations, begin analysis of data. |
|
|
Term
|
Definition
| Study: Set aside time to analyze the data and study the results. Complete analysis of data, compare data to predictions, summarize what was learned. |
|
|
Term
|
Definition
| Act: Refine the change, based on what was learned from the test. What changes are to be made? Is there a next cycle? |
|
|
Term
|
Definition
Patient feedback.
Plan: test a process of giving out patient satisfaction surveys (25)
Do: Handed out 25 surveys to patients. Noticed patients had other things to do at the time, check out area was busy, check-out attendant would let patients know they can participate in the survey.
Study: Out of 25 surveys, only 8 were received back, meaning this process did not do well.
Act: We noted that patients did not want to stay and take a survey once their visit was over. We should give the survey to take home with a stamped envelope to send back in oreder to receive better results. |
|
|
Term
|
Definition
- Study of lines
- Helps us decide how much capacity we need.
- Line is called a channel
- Infinite source: people coming into the hospital
- Finite source: physician/doctor office |
|
|
Term
|
Definition
Lambda: 10 patients per hour (arrival). Lambda=10.
Patients arrive every 6 minutes. Lambda= 10. 60 mins/6 mins
Lambda= number of patients (per hour) |
|
|
Term
|
Definition
Unit of time/service time.
- Four patients served in one hour. Mu= 4
- It takes 15 mins to serve one person. Mu=4. 60 mins/15 mins |
|
|
Term
|
Definition
- 100% utilization is great financially but would make customers mad because there is too much work being utilized by one server. = longer wait times. Could hire another part-time worker but utilization will drop to 50%.
- Number being served and idle time should represent 1.
- Persons in the system= waiting for service and/or being served. |
|
|
Term
|
Definition
- Usually for a %
- Portray the contribution of parts to a whole.
Use when: illustrating the distribution or composition of a single variable. Only for variables with mutually exclusive values (no cases are included in more than one category).
Avoid when: variables that have more than five categories. |
|
|
Term
|
Definition
- Used to show the relative size of different categories of a variable with each category or value of the variable represented by a bar.
- Shows performance results from a snapshot of time.
- Height of a bar represents the frequency of incidences for that category. |
|
|
Term
|
Definition
- Usually follows a bell curve
- Sometimes referred to as frequency distribution (how often)
- Show distribution of values as ranking along the x-axis.
- Used to display distributions of a variable that can be separated into rankings.
- Bars should touch each other except when no cases fall into an interval along the x-axis.
- They show the central tendency and variability of a data set.
- Used to quickly and easily illustrate the distribution of performance data. |
|
|
Term
|
Definition
- Tool for analyzing relationships between two variables.
- Examine theories about cause and effect relationships
- Show one of five positive correlations between the two variables.
- strong positive: value of x & y axes both increase
- strong negative: y axis decreases while x axis increases
- Possible positive: y axis increases slightly while the x axis increases.
- Possible negative: y axis decreases slightly while x axis increases.
- No correlation: no connections are evident between the two variables.
* correlation is key: no causation |
|
|
Term
| History of Quality Improvement: Types of Production Methods |
|
Definition
- Craftsman: high level of skill required, years of internships and apprenticeships required, less efficient and higher quality. Ex: rolls royce's hand stitched steering wheel.
- Assembly Line: little to no skill required (everyone is replaceable), one person performs a single task, more efficient and lower quality. Ex: For Motor Company Factory.
- Toyota Quality Assurance Method: assurance (do it right now even if it takes longer, instead of fixing it later), just in time, what/when/where is needed in the required amount, incentives for employee's to innovate ways to improve quality. |
|
|
Term
| History of Quality Improvement: Timeline (industrial quality revolution) |
|
Definition
- 1920's: Shewart introduced statistical quality control (measuring processes and applying stats to them).
- 1950's: Deming and Juran introduced quality control and quality management philosophies to Japan. (Juran quality trilogy: Quality planning- define customers and how to meet their needs, Quality control- keeps processes well, Quality improvement- lean, optimize, refine and adapt).
- 1960's: Ishikawa introduced the importance of a bottom-up approach to quality improvement (people at all levels need to be concerned about quality in the company).
- 1970's: Many US companies started losing market share to global competitors (push for business excellence).
- 1987: US business and quality leaders studied what high-performing businesses were doing better or differently (resulted in the development of criteria to support the Baldrige National Quality Program).
- 1951: ACS program transitions to the Joint Commission on accreditation of hospitals
- 1965: Medicare program (focused primarily on structural details and on underperforming hospitals and physicians).
- 1980: Joint commission adopted quality assurance standard (identify and assess important or potential problems and concerns with patient care; implement actions designed to eliminate the problems and monitor to ensure that desired results are achieved and sustained)
- 1980's: Medicare program added external quality oversight by peer review organizations |
|
|
Term
| History of Quality Improvement: Where are we today? |
|
Definition
- Regulations and accreditation standards affecting quality management are incorporating Baldrige core values and statistical thinking.
- Science of quality improvement and high reliability used in industrial and service industries is increasingly applied to the healthcare delivery system. |
|
|
Term
| History of Quality Improvement: Baldrige Core Values (11) |
|
Definition
| visionary leadership, patient-focused excellence, organizational and personal training, valuing staff and partners, agility (adjust to change), focus on failure, social responsibility and community health, managing for innovation, managing by fact, systems perspective, focus on results and creating value. |
|
|
Term
| Lean Principles and Tools: Lean vs. Six Sigma |
|
Definition
- Lean: improve process flow, eliminate non-value added actions (overproduction, waiting times, costs of quality issues). Reduce waste and increase efficiency, add value for customers, Quality Assurance.
- Six Sigma: Reduce process variation (errors and defects), reduce variation, decrease errors, improve customer satisfaction, strategic advantage (define, measure, analyze, improve, control)
*Both seek to eliminate waste and create the most efficient system possible.
*Six sigma is more of a statistical approach and about process flow. |
|
|
Term
| Lean Principles and Tools: Lean Pull Process |
|
Definition
Pull: demand drives production.
Inventory: Free capital, space (warehousing), wages (monitoring stock and reorders), avoiding obsolescence (unwanted or unused services), no stockouts (shortage of supplies)
- Ex: Kanban Cards
*Patient does not move until next step is ready. |
|
|
Term
|
Definition
- C-Charts: Measure quality defects in healthcare (bad occurrences); something that can be counted.
- P-Charts: Measure % of defects/# of samples (months). Assumptions: raw data is dichotomous (one or the other), outcome is measured over time, observations are independent (probability of one thing is not related to the other), predictions of risk are accurate, sample is representative of population. Step 1: calculate p hat. defects/# observed. Step 2: calculate control limits. |
|
|
Term
|
Definition
- The past has some relevance to the future
- Errors will occur. Predictability is not guaranteed
- Aggregate tends to be more accurate than individual
- Near term more accurate than long-term
- Forecasting anticipates future demand and plans the system.
- Good forecasts are timely, reliable, accurate, meaningful, user friendly |
|
|
Term
|
Definition
1. Identify goal
2. Estimate time horizon
3. Select technique
4. Conduct forecast
5. Select monitor |
|
|
Term
| Forecasting: Basics (types) |
|
Definition
Naive: 100 patients this week=100 patients next week.
Naive seasonality: 100 patients this October=100 patients next October.
Time Series: measurements at intervals
Delphi method: expertise to determine forecast, peoples opinions about what is going to happen and formulating a forecast. Not good because it is not quantifiable since it relies on peoples opinions.
Moving averages: average of recent data, 3 year moving average. Use a scatter plot
Weighted moving average: allows specific time periods to be more "influential", sum of weights must = 1, generally recent data will have higher weights.
Single exponential smoothing: accounts for errors in previous forecast. New forecast=old forecast+ (actual-old). Not good to use when there is a trend. |
|
|
