Presentation is loading. Please wait.

Presentation is loading. Please wait.

Quality Improvement and Patient Safety: Part 1: What is a quality chasm and why do we care??? Part II: Making Mistakes; the why’s and what if’s Part III-IV:

Similar presentations


Presentation on theme: "Quality Improvement and Patient Safety: Part 1: What is a quality chasm and why do we care??? Part II: Making Mistakes; the why’s and what if’s Part III-IV:"— Presentation transcript:

1 Quality Improvement and Patient Safety: Part 1: What is a quality chasm and why do we care??? Part II: Making Mistakes; the why’s and what if’s Part III-IV: You and I Can Do QI Danielle Scheurer, MD, MSCR Chief Quality Officer

2 Learning objectives: Understand what quality and QI are Understand why and how QI evolved Understand why QI is important MUSC and DOM QI structure

3 What is quality? Meeting the needs and exceeding the expectations of those we serve Delivering all and only the care that the patient and family needs

4 What is improvement? Reducing variation in care processes Shifting and tightening of the bell curve A change in the design of care Ultimate goal is outcome improvement

5 Quality Improvement worse better worse better Quality After Before Quality worse better Quality After Quality Quality Assurance Quality Improvement Bell Curve: Patient Population Tail  better Improvement and standardization in processes reduces variation (narrows the curve) and raises quality of care for all (shifts entire curve toward better care). This radical change is what defines Quality Improvement.

6 What put quality on the “front burner” Libby Zion: 1984 Lewis Blackman: 2000 Josie King: 2001 Young Healthy Smart parents 97,999 low profile cases

7 What put quality on the “front burner” To Err Is Human ◦ 1999 IOM Report ◦ 98,000 deaths in US hospitals a year due to preventable medical errors ◦ Due to “faulty systems, processes, and conditions that lead people to make mistakes” Crossing the quality chasm ◦ 2001 IOM report ◦ Defined 6 dimensions of healthcare quality ◦ Suggestions for improvement

8 Institute of Medicine Annual deaths in the US from: Medical errors: 44,000 – 98,000 Motor vehicle accidents: 43,458 Breast cancer: 42,297 AIDS: 16,526

9 Institute of Medicine “The nation’s health care system lacks… the capabilities to ensure that services are safe, effective, patient-centered, timely, efficient and equitable…Between the health care we have and the care we could have lies not just a gap but a chasm.” The American health system needs fundamental change

10 Dimensions of Healthcare Quality Safe: Avoid patient injuries from the care that is intended to help them Effective: Match care to science; avoid overuse of ineffective care and underuse of effective care Patient-centered: Honor the individual and respect choice Timely: Reduce waiting for both patients and providers Efficient: Reduce waste Equitable: Close racial and ethnic gaps

11 Quality Improvement: Bridges the Implementation Gap Implementation Gap Scientific understanding Patient care Progress Time

12 Why is it important? Making mistakes: It happens Transposition error rate (mixing drug doses) Commission error rate (misreading a label) Calculation error rate Omission error rate (failure to turn off a switch) Error rates under stress or danger Most serious medical errors are committed by competent caring people doing what other competent caring people would do…..Don Berwick

13 Why is QI important? Patients: Better care, fewer errors Regulation and accreditation Reputation Financial Life long learning (ACGME and MOC) Academic activity Front-line clinicians – Empowerment to change – Jobs more rewarding – Less daily frustration

14 Quality and the ACGME System based practice: Residents must demonstrate and awareness of and responsiveness to the larger context and system of health care, as well as the ability to call effectively on other resources in the system to provide optimal health care

15 System based practice: Expectations Work effectively in various health care delivery settings and systems relevant to their clinical specialty Coordinate patient care within the health care system relevant to their clinical specialty Incorporate considerations of cost awareness and risk-benefit analysis in patient care and/or population- based care as appropriate Advocate for quality patient care and optimal patient care systems Work in interprofessional teams to enhance patient safety and improve patient care quality Participate in identifying system errors and in implementing potential systems solutions

16 Quality and the ACGME Practice based learning and improvement: Residents must demonstrate the ability to investigate and evaluate their care of patients, to appraise and assimilate scientific evidence, and to continuously improve patient care based on constant self-evaluation and life-long learning. Residents are expected to develop skills and habits to be able to meet the following goals:

17 Practice based learning and improvement Identify strengths, deficiencies, and limits in one’s knowledge and expertise Let learning and improvement goals Identify and perform appropriate learning activities Systematically analyze practice, using quality improvement methods, and implement changes within the goal of practice improvement Incorporate formative evaluation feedback into daily practice Locate, appraise, and assimilate evidence from scientific studies related to their patients’ health problems Use information technology to optimize learning Participate in the education of patients, families, students, residents, and other health professionals

18 Quality in Life Long Learning American Board of Internal Medicine “Maintenance of Certification” requires evidence of participation in quality / process improvement Subspecialty certification requires QI

19 Career No longer “soft science” Peer reviewed publications Considered “translational research” Subject to IRB for dissemination

20

21 QI and ACGME Medicine Peds, Optho, Neuro, Derm, Radiology, Anesthesia ……

22 Resources Curriculum Early project development Data resources Statistical help

23 PI Project at MUSC

24 Quality Improvement at MUSC IMPROVE structure for QI Lean and Six Sigma management philosophies, along with strong leadership and willingness to innovate, to improve quality

25 Quality Philosophy: Lean Management philosophy derived from Toyota Production System which focused on the elimination of waste Waste is anything that adds to a process without adding value “non-value added”

26 What is “Waste”: 8 types Waiting: Patients, doctors, staff Overproduction: 3 admits in 20 minutes to the same person Rework: Medication reconciliation Motion: Looking for charts Processing: Several people doing the same steps to discharge a patient Intellect: Doing things “below your pay grade” Inventory: Wasted “stuff” laying around Transportation: Patients, materials

27 Quality Philosophy: Six Sigma Management philosophy that seeks to reduce variation and defects by applying a highly structured, systematic approach to operational success ◦ Asthmatics should all get beta-agonists ◦ Sickle cells admits should all get IVF Initially developed by Motorola

28 IMPROVE Identify the problem Measure the impact Problem analysis Remedy the critical issues Operationalize the solutions Validate the improvement Evaluate over time

29 Identify the problem What is the problem? Who identified it? When was it identified? When and where is it occurring? Pick something that matters to you, and state WHY it matters (how does it fit into the domains of quality) Who else cares about the problem (who are the stakeholders?)

30 Measure the impact How is the data collected (survey, administrative data, chart review)? Is it valid/accurate? Is it a manual process or automated? Is there a clear definition of the outcome (or can it be interpreted different ways)? Who is going to measure? Can you sample?

31 Problem Analysis: What is causing the problem? Time of day, day of week Department specific / system wide Inefficient staffing (numbers or skill set) Poor communication Inadequate process or policy Lack of controls to keep the problem from occurring Poor individual performance (usually not the only issue) Pick an appropriate process analysis tool to further analyze the problem/process

32 Remedy the critical issues Pick a remedy based on the process analysis What will the group try to improve? What are the barriers? What evidence is there that it will have an impact (has someone tried and succeeded or failed)? How “reliable” is the intervention? Do you need more than one intervention to make it nearly impossible to recur?

33 Remedies (in order of reliability) Education Reminders Checklists Order sets Protocols Pathways Templates “Hard stop” order entry

34 Operationalize How are you going to make it work? How will the barriers be removed? What assistance is required from senior leadership? What is the plan to roll out and implement solutions?

35 Validate How will we know we made a difference? What are you measuring? How often are you measuring it? Is the measure meaningful? Are you measuring “unintended consequences”?

36 Evaluate How to sustain the improvement? Who is responsible for monitoring and measuring over time? What is the plan to react if the measures slip? How will future staff be made aware of the new process?

37 What is a Safety Culture? Not focused individual blame Safe system to disclose (potential) issues without repercussion System approach to proactive error reduction Workers are involved in the work process and process re-design Workers are empowered to make change

38 How to create a safety culture No blame problem analyses Anonymous safety reporting (eg PSN) Providing feedback to front line staff of events and reporting Frontline provider engagement in changing the system

39 Making mistakes The why’s and what if’s Danielle Scheurer, MD, MSCR Chief Quality Officer

40 Learning objectives: Understand basic definitions in patient safety (errors, adverse events, negligence) Understand fundamentals of “safety culture” and “system approach” to safety Understand most common types of healthcare errors and how to mitigate Understand the role of disclosure, apology, and renumeration in safety events

41 What is an Error? Failure of a planned action to be completed as intended OR the use of a wrong plan to achieve an aim - To Err is Human, Institute of Medicine 2000

42 Types of errors: Latent System defects that make active errors more likely to happen Ex: Poor design of IT data display making it easy for clinicians to misunderstand ◦ Nursing understanding of heparin protocols Inadequate staffing making people “rush” or routinely “multi-task” Difficult to measure; may exist for years before effect is seen

43 Types of errors: Active Occur at the frontline provider level Usually easier to measure; negative outcome almost immediately apparent ◦ Nurse hangs the wrong medication ◦ Doctor orders medication intended for another patient ◦ PICC line inserted into the wrong patient ◦ Doctor amputates the wrong leg

44 Active versus latent errors A “safety culture” system: Proactively looks for latent errors all the time Retroactively looks for latent errors when an active error has occurred

45 What is an Adverse Event? An injury that is caused by medical management rather than the patient’s underlying condition Preventable adverse event = adverse event due to error

46 Errors Adverse Events Preventable Adverse Events Near Misses Negligence

47 The points of intersection Not all errors result in an adverse event (“near miss”) Ex: double dose lasix with no effect to the patient Not all adverse events are the result of an error Ex: hyperglycemia from steroids Of errors/adverse events that are preventable, only a small portion are due to “negligent” active error

48 Patient Safety Freedom from accidental injury Involves the establishment of systems to decrease all errors, and intercepting them when they occur

49 What is a Safety Culture? Not focused individual blame Safe system to disclose (potential) errors without repercussion System approach to proactive error reduction Workers are involved in the work process

50 How to create a safety culture No blame M&M conferences or RCAs Simulating possible adverse events (eg FMEA) Anonymous safety reporting system (eg PSN) Providing feedback to front line staff of events and reporting

51 System Approach to Error: Fundamentals Analyze the system for contributing factors that allow errors to occur (majority of errors due to faulty systems) Individual blaming will not prevent future errors from occurring Individuals for the most part are trying to do their best; they do not come to work with the intent to harm

52 Making mistakes: It happens Transposition error rate (mixing drug doses) Commission error rate (misreading a label) Calculation error rate Omission error rate (forget) Error rates under stress or danger X

53 A system approach “We cannot change the human condition, but we can change the conditions under which humans work” Reason; BMJ 2000;320:

54 Human Factors Engineering “The study of the interrelationships between humans, the tools they use and the environment in which they live and work”

55

56 Factors that contribute to error Environment: Distractions, interruptions, noise, ergonomics, clutter, lighting ◦ Forgot to enter order due to multiple pages/competing priorities ◦ Lost handwritten sign out sheet and forgot to check a lab ◦ Discharge paperwork incomplete due to inability to find all needed papers

57 Factors that contribute to error Human: Stress, fatigue, emotions (boredom, frustration) communications ◦ Feel asleep and did not check a lab ◦ Failure to clarify sign out questions due to discomfort with confrontation ◦ Transcription error in filling out discharge med list due to stress or fatigue

58 Factors that contribute to error Patient: language, compliance, fear, disease acuity, knowledge deficit ◦ Patient neglects to mention a pertinent review of symptoms due to fear or embarrassment of what it would mean ◦ Patient neglects to take meds correctly due to finances, lack of understanding, low literacy ◦ Patient does not understand instructions due to language barrier

59 Factors that contribute to error Equipment: design, variability, quality, labeling, training, availability ◦ Patient gets wrong heparin dose due to poor nurse training on how to program smart pump, multiple heparin doses in drug library, or no availability of a pump ◦ Patient gets wrong chemotherapy dose due to confusing labeling during packaging ◦ Hand hygiene does not occur due to unavailability of purell/soap

60 Factors that contribute to error Systems: staff level, staff experience, supervision, handoffs, workflow ◦ CLA-BSI occurs due to low staffing and single provider putting in line without checklist or help with sterile procedure ◦ Med rec not completed because it is out of workflow (does not connect to CPOE or inpatient documentation) ◦ Patient not seen upon transfer out of ICU due to inadequate communication from ICU to floor team

61 Layers of defense to reduce errors

62 Most common types of errors in healthcare Medication errors Procedural errors Communication errors Diagnostic errors

63 Medication Errors ADE: ◦ Increases LOS by 8-12 days ◦ Increases cost $16,000 to $24,000 ◦ Most commonly due to wrong dose (not time, route, drug, or allergy errors) pADE: ◦ Most commonly occurs at physician ordering stage (not administration or dispensing) Near miss: Caught before given

64 Mitigating Medication Errors Order stage: ◦ CPOE with decision aids for dose, route, interval, indication Prep stage: ◦ Double check with pharmacy for drug/dose appropriateness ◦ Remove open floor stocks of high risk medications (heparin, insulin)

65 Mitigating Medication Errors Administration stage:  Bar coding to match patient with medication dose, time, and route  Teach back technique with the patient/family during medication administration

66 Procedural errors Wrong patient Wrong site Wrong/Bad equipment

67 Mitigating procedural errors Procedural “time- out” with checklists Team training Culture of safety with frontline empowerment Teach back technique with the patient for procedural marking (prior to sedation)

68 Mitigating procedural errors Procedural simulation training Consistent equipment stocking

69 Communication errors Illegible handwriting Inadequate handoffs Poor or no communication among or between teams Poor, no, or inappropriate communication to the patients

70 Mitigating communication errors Template EMR documentation Multi-disciplinary team training and rounding Standardize template, time, and place for handoffs (without noise, distraction)

71 Mitigating communication errors Use read back techniques for critical information Use SBAR technique (or other) to communicate new patient findings

72 Diagnostic errors Delayed or missed diagnosis which affects patient outcome or leads to additional complications / procedures Account for about 1/3 of errors Associated with proportionally higher morbidity than other errors Can be due to heuristics, most commonly ◦ Pattern recognition ◦ Algorithmic thinking

73 Pattern recognition: What do you see?

74 Pattern Recognition: A heuristic (shortcut) You see it, you know it Relies on experience Requires a well defined prototype Is quick and “easy” and used commonly Prone to representative error and premature closure

75 Algorithmic Thinking Solute loss (eg vomiting) SIADH Cirrhosis, CHF, nephrotic syndrome Hyponatremia HypovolemicNormovolemicHypervolemic

76 Avoiding diagnostic error Slow down ◦ Take a diagnostic time out ◦ Always make a differential diagnosis ◦ Ask what does not fit, and could it be more than one diagnosis

77 Avoiding diagnostic error Always consider the worst it could be Focus on the history and physical exam ◦ Listen to the patient ◦ Reduce reliance on testing

78 Avoiding diagnostic errors Understand Bayesian theory ◦ Use a mathematical approach to decision making ◦ Use pre and post-test probability to know the value of diagnostic testing

79 Avoiding diagnostic errors Acknowledge how the patient makes you feel ◦ Emotions (good and bad) can affect how you test and treat the patient

80 Medical negligence Failure to meet the standard of practice of an average qualified physician practicing the specialty in question Physician practicing in a system not grossly “latent error prone” that breaches standard practice

81 Disclosure: part of patient safety Errors and near misses have to be transparent ◦ To providers ◦ To patients ◦ To the system Disclosure with compensation reduces claims, lawsuits, and legal costs Kachalia A. Ann Intern Med 2010;153:213.

82 What physician think

83 What Patients Want Jargon-free statement that error occurred Description of the error Why it happened Implications of the error for their health and how to deal with the consequences 83 Gallagher TH, et al. JAMA. 2003:289:

84 What Patients Want Outline of steps that will be taken to prevent future errors An apology from the health care worker 84 Gallagher TH, et al. JAMA. 2003:289:

85 Why an apology?

86 Barriers to Disclosure for Physicians Fear of litigation ◦ Unlikely to apologize due to concern about consequences of admitting fault Discomfort with discussing such issues ◦ Physician may choose words carefully to avoid explicitly stating that an error occurred Concern that information may harm patient ◦ Belief that disclosure may impact patient’s trust in the physician 86 Gallagher TH, et al. JAMA. 2003:289:1001-7; Robinson AR, et al. Arch Intern Med. 2002:162: ; Wu AW, et al. JAMA. 1991;265:

87 Errors happen: 3 part plan RECOGNIZE the mistake MITIGATE harm associated with error DISCLOSE the error

88 MUSC approach to apology: DO Take care of the patient Call Risk management Call your attending Be open and honest

89 Disclosure process: If there is a change in plan of care: Attending must disclose If there is not a change in plan of care: Attending must be informed, but resident can disclose Always call risk management first

90 How to disclose Apologize Explain outcome, current condition, treatments Provide support and empathy Offer family meeting Be accessible Give contact info Use terms negligence or accident Blame or discuss fault Admit or speculate about liability Discuss peer review or root cause analysis DODO NOT

91 How to document Objective clinical information Medical response Follow up plans If late entry, mark “late entry” with date of time of event and entry Alter prior documentation Mention risk management, occurrence reporting, sentinel event DODO NOT

92 Steps for Risk Patient follow up plan after analysis of what happened Financial plan Attending follow up

93

94

95 Quality Improvement: Part 3-4 You and I Can do QI Project Design and Publication Danielle Scheurer, MD, MSCR, SFHM Medical Director of Quality Medical University of South Carolina

96 Learning objectives: Understand the fundamentals of forming a team and team meetings Give an example of a goal and an aim Give examples of types of measurement Identify some ways to evaluate a process Understand benefits of PDSA

97 Team Members Champion: Respected, able to surmount barriers, external voice of the project (mentor, attending)

98 Team Members Team Facilitator: QI expert, helps with design and operations (optional but nice) Process owner: Schedules meetings, sets the agenda, records activity, reports findings. “Owns” the success of the project.

99 Team Meetings All members and opinions are equal Team members will speak freely and in turn, no side bar conversations, one at a time Cells phones and pagers on silent alert Discuss/analyze/attack problems, not people Everyone gets their homework done Silence (or absence) equals agreement Meetings start / end on time Get creative on meeting times, use conference calling

100 IMPROVE Identify the problem Measure the problem Problem analysis Remedy the critical issues Operationalize the solutions Validate the improvement Evaluate over time

101 Identify the problem What is the problem? Who identified it? When was it identified? When and where is it occurring? Pick something that matters to you, and state WHY it matters (how does it fit into the domains of quality) Who else cares about the problem (who are the stakeholders?)

102 Measure How to get data ◦ (survey, administrative data, chart review)? Is it valid/accurate? Is it a manual process or automated? Is there a clear definition of the outcome (or can it be interpreted different ways)? Who is going to measure? Can you sample?

103 Problem Analysis: What is causing the problem? Time of day, day of week Department specific / system wide Inefficient staffing (numbers or skill set) Poor communication Inadequate process or policy Lack of controls to keep the problem from occurring Poor individual performance (usually not the only issue) Pick an appropriate process analysis tool to further analyze the problem/process

104 Set Goal Goals are general statements of what you are trying to achieve – Reduce CA-UTI

105 Set Aims Aims are SMART – Specific: Reduce CA-UTI on 8E – Measurable: Current 8/1000 catheter days – Achievable: Reduce to 5/1000 catheter days – Realistic: Depends on HOW to achieve it – Time delimited: In 6 months

106 Examples of Aims > 75% of CHF patients on 8E will be discharged with instructions within 6 mo Decrease all cause mortality rates in COPD from 10% to 5% by Jan 2012 Reduce unplanned extubations by 50% within 3 months of intervention 106

107 Types of Measurement Process measures: The process of delivering care ◦ HbA1c for diabetics ◦ Completion of an order set for asthmatics Outcome measures: The actual end point ◦ Death ◦ Readmission rates Balancing measures: What are the other (sometimes adverse or unintended) consequences of our intervention ◦ Time to complete work ◦ Longer length of stay

108 Measurement: How will we know that a change is an improvement? Pick measures that are related to aims Develop operational definitions Determine patient eligibility, denominator, numerator Keep measures simple and practical 108

109 Examples of Measures Mean time from initiation of discharge process to patient walking into elevator (in minutes) ◦ [Process measure] % PNA patients readmitted with 30 days ◦ [Outcome measure] PNA length of stay ◦ [balancing measure] 109

110 Principles of measurement Seek usefulness, not perfection (your goal is improvement, not measurement) Integrate measurement into the daily routine (instead of waiting 2 months for administrative data) Use qualitative and quantitative data (ask nurses how it is going?) Use sampling (make your measurement easy, not hard) Plot data over time (to observe trends and patterns) Use balancing measures (measure anticipated unintended consequences, such as time or cost) Measure both process and outcome measures (% use of a discharge order set, and % discharged on intended medication)

111 Measuring processes: ◦ Cause and effect diagrams ◦ Process maps ◦ Swim Lane Maps ◦ Decision flow diagrams ◦ Spaghetti diagrams ◦ Cycle time charts Basic 2 types of data analysis tools ◦ Pareto chart ◦ Run chart

112 Process evaluation: Processes are complicated, involving multiple steps and decision branch points If a process is flawed, inefficient, or resulting in an undesirable outcome, the process should be systematically evaluated How to evaluate a process: ◦ Cause and effect diagrams ◦ Process map ◦ Swim Lane map ◦ Decision Flow Diagram ◦ Spaghetti diagrams ◦ Cycle time charts

113 Cause and Effect Diagram Fishbone “Ishikawa” diagram Graphically displays possible factors leading to an event or process Used to organize ideas during brainstorming sessions Helps to logically and systematically evaluate a process, and decide which steps are flawed, inefficient, or need repair

114 Cause-and-Effect Diagram Physiologic Factors Pharmocologic Factors Drug Administration Errors Ordering Errors Transcribing Spelling Pharmacokinetics Renal Dilution Time Nurse Route Rate ADE Nurse Physician Pharmacist Physician Pharmacy Nurse/Clerk Pharmacist Patient Physician Dietician Patient Wrong Drug Dose Scheduling Dosage Route Past Allergic Reaction Absorption Weight Age Gender Electrolyte Hepatic Race Pharmacodyamics Expected Drug/Drug Unforeseen Drug/Food Drug/Lab Cognitive Psychiatric Compliance Patient Errors Order Missed Place outcome here Example: Adverse Drug Events (ADE)

115 “Macro” Process Map Displays the number of steps needed to successfully implement a process Graphically displays the “current state” at a “high level” Each step is displayed, which can later be broken down into smaller steps “micro level” The process, and each point, can be further investigated based on where the failures are occurring

116 Macro Process Map Patient needs a procedure INR checked and elevated FFP ordered and given

117 Swim Lane Map Shows simultaneous workflows of different disciplines Can see the interaction of disciplines Helps to facilitate better communication among disciplines in complicated processes

118

119 Decision Flow Diagram Displays the steps of a process with decision branch points Creates transparency of a process regardless of complexity Can identify areas of improvement ◦ How the process is supposed to work ◦ How the process is actually working ◦ Variability in the process Circle (input/output), diamond (decision branch), square (task to do)

120 Decision Flow Diagram Lamp not working Plugged in? Bulb burned out? Plug it in Put in new bulb Wor king? Buy new lamp Stop Y Y N N Y Y N

121

122 Spaghetti diagram Graphically displays motion (and usually waste) of a process (or set of processes) Can be effective for using before-after an intervention, although less “rigorous” an outcome measure Can be used in conjunction with time motion studies, pedometer readings, etc

123 Outpatient clinic

124 Cycle times Graphic depiction of what processes are occurring, in what order, and how long each “batch” is taking Anytime the “batch” times are uneven, there is waste built in the process Efficient processes have “load balancing” Can be used to identify waste within a process, then used to design a solution, which can be re-measured

125 ED process: cycle times

126 Get more granular; see if you can load balance work

127 Once you have collected data: Data Analysis Data can be displayed (and analyzed) in a number of ways, the most common (and simple) of which are: ◦ Pareto Chart (Bar graph) ◦ Run Chart

128 Pareto Chart (Bar graph) Graphically displays relative frequency of events Used to focus choices, by designating the weight of each category Can help teams prioritize which processes to focus on first Based on this graphic depiction of the causes of an adverse event, might focus on the top 3 contributing causes first

129 Causes Pareto Chart % Contributing Causes Contributing to Adverse Drug Events

130 Run Charts Graphically displays information over time The horizontal axis usually has a mean performance The vertical axis has interventions placed to determine if the interventions are affecting the outcome

131 Run Chart

132 Now it is time to implement the change Start on a small scale (a clinic, a unit, a doctor) Improve incrementally Learn through action Test your changes Assess their effect Re-work the changes and do it again and again and again… PDSA cycles: PLAN, DO, STUDY, ACT

133 Value of PDSA cycles Allows “failures” without impacting large scale performance (“fail fast”) Adapts to changing environments Evaluates costs and unintended consequences of the change Empowers front line users to experience the change on a small scale Minimizes resistance for larger implementation

134 Bottom line with PDSA Common sense approach to change and improvement It is not complicated Manageable ◦ big ideas tested on a small scale Cycles happen quickly ◦ allows for successive cycles and sustainable change Guides the test of a change to determine if the change is an improvement

135 Changes that result in Improvement P DS A P DS A P DS A P DS A Time Source: D. Berwick Big Idea

136 IMPROVE: Next week Identify the problem Measure the impact Problem analysis Remedy the critical issues Operationalize the solutions Validate the improvement Evaluate over time

137 Remedy the critical issues Pick a remedy based on the process analysis What will the group try to improve? What are the barriers? What evidence is there that it will have an impact (has someone tried and succeeded or failed)? How “reliable” is the intervention? Do you need more than one intervention to make it nearly impossible to recur?

138 Operationalize How are you going to make it work? How will the barriers be removed? What assistance is required from senior leadership? What is the plan to roll out and implement solutions?

139 Validate How will we know we made a difference? What are you measuring? How often are you measuring it? Is the measure meaningful? Are you measuring “unintended consequences”?

140 Evaluate How to sustain the improvement? Who will monitor / measure over time? What is the plan if the measures slip? How will future staff be made aware of the new process?

141 Remedies At least one needs to be “highly reliable” Usually need several to be effective and sustainable ◦ Admission / Discharge orders ◦ Order entry “hard stops” ◦ Decision aids / Protocols ◦ Educational material ◦ Pre-populated documentation in EMR ◦ Standing orders / Default pathways

142 Remedies (in order of reliability) Education Reminders Checklists Order sets Protocols Pathways Templates “Hard stop” order entry

143 What is “Reliability” Failure free operation over time You get the results you intended to get Inverse of the system’s failure rate

144 Making Systems Safer: Way to Redesign Systems Elimination Standardization Constraints and forcing functions Redundancy Differentiation Confirmation Simplification Reminders Detection Mitigation User-centered design 144

145 Organizational Reliability Aviation and nuclear power: Failure rate is 1 in 1,000,000 (“Six Sigma”) Medical processes: we routinely “fail” ◦ Hand hygiene: 8 in 10 ◦ Patient understanding of medication: 1 in 3 ◦ VTE prophylaxis in inpatients: 9 in 10

146 What is “High Reliability” High reliability system Known to be complex and dangerous, but safe with infrequent error rates Has a preoccupation with failure Defers decisions to expertise (not seniority) Routinely (and proactively) evaluates what could go wrong, instead of what did go wrong Seeks to standardize processes; doing the “right” thing is easier than doing the “wrong” thing DOES NOT rely on vigilance or memory alone

147 Not Highly Reliable

148 Highly Reliable

149 Why do we care? Ways to think about designing interventions that will be ◦ EFFECTIVE ◦ SUSTAINABLE

150 Elimination Eliminate some choices on purpose Reduce # procedural trays available ◦ Combine with simulation to ensure understanding and competency Reduce # medications available ◦ Combine with CPOE and formulary aids 150

151 Elimination Remove high risk medications from floor stock (concentrated KCl, insulin vials) Do not allow verbal orders Remove razors from all procedural areas 151

152 152 Standardization Prescription writing ◦ 0.1 mg but not 1.0 mg ◦ Approved abbreviations only (no “QD”) ◦ Spell out units: “micrograms”, not “µg” Metric units only (kg, cm, Celsius) ◦ Standardize all thermometers, bed weight, order sets, patient instructions

153 153 Standardization Standardize equipment ◦ Pumps ◦ Ventilators ◦ Keyboard Standardize documentation ◦ Flow sheets ◦ Progress notes ◦ Discharge summaries

154 5S Standardization Workplace organization methodology Sort: Keep essential items, discard others Straighten: Everything has it’s place Sweep/Shine: Visual cue for mess Standardize: All units identical Sustain: Avoid entropy Creates safer environments

155

156 Constraints / Forcing Functions CPOE Orders with pre-printed boxes checked Metric system for orders Scan armband to allow medication release Not allowing blood release from blood bank until 2 blood types verified 156

157 Redundancy 2 nurse verification for chemotherapy Pharmacist verifies admission medications from physician medication history Pharmacist “signs off” on all new medication orders 157

158 Redundancy 2 patient identifiers before blood drawn Critical lab read back technique Patient weight on every order page Team contacted twice for ICU transfers 158

159 Differentiation

160 Confirmation Allowing for question time during sign out Read-back of verbal orders “I’m clear, you’re clear, we are all clear” during codes Procedural time out 160

161 Simplification 1-2 medications from each category available for ordering (ACE, BB) Reduce number of places information is stored (vitals signs, allergies) Universal influenza vaccination for all inpatients

162 Simplification Make things predictable and easy to find (oxygen tubing, a central line kit, LP tray) Have central line kits with only necessary equipment included Reduce med rec from a 2 to a 1 step process

163 Reminders: Avoid relying on memory Checklists: first developed in mid 1930’s by the Army Air Corps in response to increasing complexity of flying an airplane. ◦ Resulted in the adoption of the B-17, which helped to turn the course of WWII Adopted as a way to reduce central line associated infections, VAP, wrong site surgery, etc, etc, etc 163

164 Detection Create a system so that errors are detectable in real time Bar coding / scanning meds, makes it easy to recognize you are about to give the wrong med to the wrong patient

165 Detection Hard stop in CPOE for allergy and drug interactions Sponge count at the end of surgeries Xrays at the end of all trauma cases

166 Mitigation When something goes wrong, make the impact as innocuous as possible Standard reporting for events Immediate RCA to reduce harm associated with error and recurrence Give medication antidote if available

167 User-centered designs Creating a work space that enables us ◦ To make safe and sound decisions ◦ To implement those decisions into a seamless care plan

168 Not user centered

169

170

171

172

173

174 QI/PS Resources IHI (http://www.ihi.org/IHI/Topics/PatientSafety) AHRQ (http://www.ahrq.gov/qual)http://www.ahrq.gov/qual AHRQ 2009 quality report (http://www.ahrq.gov/qual/nhqr09/nhqr09.pdf) JC (http://www.jointcommission.org/Library) NPSF (http://www.npsf.org/rc/mp) CDC (http://www.cdc.gov/ncidod/dhqp/index.html) IOM (http://www.iom.edu) SHM (http://www.hospitalmedicine.org) HRSA (http://www.hrsa.gov/quality) SQUIRE guidelines/resources (http://squire-statement.org/)http://squire-statement.org/ AHA HRET (http://www.hret.org/hret/publications/qualitypubs.html)http://www.hret.org/hret/publications/qualitypubs.html American Society of Quality (http://www.asq.org/learn-about-quality/cause-analysis- tools/overview/overview.html) QI/PS Journals International Journal for Quality in Health Care International Journal of health care quality assurance Joint Commission journal on quality and patient safety Journal of health care quality

175 Enhanced Reading Davidoff F, Batalden P, Stevens D, Ogrinc G, Mooney S for the SQUIRE development group. Publication guidelines for quality improvement studies in health care: evolution of the SQUIRE project. JGIM 2008;23:2125. Nolan and Berwick. All-or-none measurement raises the bar on performance. JAMA 2006;295:1168.

176 Enhanced Reading ovementMethods/HowToImprove ovementMethods/HowToImprove es/1999/To-Err-is- Human/To%20Err%20is%20Human% %20%20report%20brief.pdf es/1999/To-Err-is- Human/To%20Err%20is%20Human% %20%20report%20brief.pdf IHI open school ◦ QI 101: Fundamentals of improvement QI 101: Fundamentals of improvement ◦ PS 101: Introduction to patient safety PS 101: Introduction to patient safety

177 Writing an abstract for a scientific meeting:

178 Learning objectives Recognize the differences between abstracts for manuscripts and for meetings Compare Traditional vs QI abstracts Understand key steps in writing a good abstract Identify specific strengths & common problems in abstracts submitted for presentation at scientific meetings.

179 Outline Why write a great abstract? Review the anatomy of an abstract Review common errors in abstract writing Small groups to critique previously presented abstracts Large groups to discuss findings and address specific comments

180 Why Write “Abstracts for Scientific Meetings” Chaotic ideas find form and substance through drafting Promotes original thought Sets goals for project completion

181 The Project Life Cycle Project Idea Preliminary Data Abstract for Scientific Meeting Peer- Reviewed Publication Advance Science Obtain Funding and Career Advancement

182 Purpose of an Abstract for a Scientific Meeting Flash compared to detail Limit data presented but also present most important data ◦ What data supports your conclusions directly?

183 Guidelines for Abstract Writing

184 Title Concise Direct No abbreviations

185 Introduction Why this study? Concisely state the goals and rationale (i.e., describe why the work was done in 1 or 2 sentences) 10% to 15% of total length

186 Methods What was done? ◦ Describe the population studied, techniques used, and how the data were analyzed ◦ Definitions 30% to 35% of total length

187 Results What was found? ◦ Summarize only the important findings Tables & graphs can be used in abstracts written for meetings as a means to emphasize the results ◦ A table or graph is not a substitute for results but rather a means by which to efficiency display the results 35% to 40% of total length

188 Discussion The Answer!! ◦ Most important things learned ◦ OK to state implications of findings but don’t extrapolate too much ◦ Keep conclusions within scope of data investigated 10% to 20% of total length

189 QI vs. Traditional Manuscripts Traditional Manuscript Completely defined process Generalizability Methods section Describe design Details about all of process Results section Present summary measures Adjust for variability Tables and Figures Discuss: Interpret data and describe impact QI Manuscript Evolving process Local context Methods section Describe context/personnel Detail about initial process Results section Describe process changes Show all the data/variability “Run” and SPC charts Discuss: implications of outcome data and changes in process Davidoff F, Batalden P, Qual Saf Health Care 2005

190 The Writing Process: Drafts Create an outline or free-writing under headings A first draft will likely be disappointing ◦ You can’t revise until you’ve written

191 The Writing Process: Drafts Obtain feedback from peers ◦ Encourage them to provide an honest and constructive critique ◦ “I want this to be a good abstract so please don’t worry about me being offended” More is better ◦ Solicit feedback from multiple people from diverse disciplines ◦ Different reviewers focus on different areas & content (e.g., research design, statistical analysis, or writing style)

192 The Writing Process: Drafts Provide feedback to peers when asked Participate in group feedback sessions- you get to hear a variety of viewpoints The more abstracts you review, the easier it is to write your own

193 Writing Styles & Hints No jargon Only use standard, well-accepted abbreviations and symbols Shorter is better, always ◦ I have made this letter longer than usual because I lack the time to make it shorter (Blaise Pascal, Provincial Letters, XVI. 1657)

194 Common Problems Lack of sentence clarity (i.e., the message is not clear) ◦ Question omitted ◦ Question vaguely stated ◦ Answer is not stated (only implications stated)

195 Common Problems Poor organization ◦ No clear ordering of sections ◦ Mixing content (e.g., giving results in the methods section)

196 Common Problems Lack of coherence ◦ Too many abbreviations ◦ Use of “respectively”  Cure was achieved in patients with aa, bb, cc, dd, and ee in 0.5%, 0.5%, 0.6%, 0.7%, and 0.8%, respectively  Better to use a table

197 Common Problems Excessive detail ◦ Reporting data for minor results ◦ Providing details for well-known techniques ◦ Providing exact data rather than odds ratio or percent change ◦ Duplicating data ◦ Including references

198 Common Problems Failure to consider non-specialist readers ◦ Underdeveloped rationale ◦ Overly technical language

199 Common Problems Make sure the entire abstract develops and concludes the main point YOU want to express ◦ Abstracts evolve as you write them  Critically review the introduction statements after finishing conclusions

200 The Project Life Cycle Project Idea Preliminary Data Abstract for Scientific Meeting Peer- Reviewed Publication Advance Science Obtain Funding and Career Advancement

201 Publication Rate Carroll AE, et al. Pediatrics 2003;112:

202 It Takes Time, Be persistent Carroll AE, et al. Pediatrics 2003;112:


Download ppt "Quality Improvement and Patient Safety: Part 1: What is a quality chasm and why do we care??? Part II: Making Mistakes; the why’s and what if’s Part III-IV:"

Similar presentations


Ads by Google