Overview of Quality ImprovementFocus on Designing Reliable Interventions: Overview of Quality Improvement Focus on Designing Reliable Interventions Greg Maynard MD, MS Professor of Clinical Medicine and Chief, Division of Hospital Medicine University of California, San Diego


Slide2: Quality Improvement: Bridging the Implementation Gap


Working harder isn’t always the answer……: Working harder isn’t always the answer……


The Evolving Culture of Medicine : The Evolving Culture of Medicine 20th Century Characteristics Autonomy Solo Practice Continuous learning Infallibility Individual Knowledge 21st Century Characteristics Teamwork & systems Group practice Continuous improvement Multidisciplinary problem solving Change Shine, KI. Acad.Med. 2002;77:91-99


How Do We Close the Gap? Essential Elements: How Do We Close the Gap? Essential Elements Institutional support and multidisciplinary teams Standardized order sets Infusion Subcutaneous which promote basal / bolus regimens Algorithms / protocols / policies Address dosing Nutritional intake Special situations: TPN, enteral tube feedings, perioperative insulin, steroids Safety issues Transitions in care and discharge planning Metrics: How will you know you’ve made a difference? Comprehensive educational program


Traditional Quality Assurance: Traditional Quality Assurance outliers


Slide7: Quality Improvement worse better worse better Quality Quality After Before


Slide8: Quality Improvement is… Focus on processes of care Reduced variation by shifting entire practice A change in the design of care Quality Improvement is NOT… Forcing people to work harder / faster / safer Traditional QA or peer review Creating order sets or protocols without monitoring use or effect


Good Teamwork is Essential: Good Teamwork is Essential


Features of a Good Team: Features of a Good Team Safe no ad hominem attacks Inclusive open to all potential contributors values diverse views; not a clique Open considers all ideas fairly Consensus seeking finds a solution all members can support


Models for Improvement: Models for Improvement In use around the globe for decades Success in many fields of endeavor Healthcare late to the game! Alternative to the usual: Predictable breakdowns in reliability leading to common problems Ignoring improvement concepts & trying the first thing that comes to mind Not measuring effectiveness of implementation outcomes or process until bad events happen…..again


Slide12: Establishing Measures Teams use quantitative measures to determine if a specific change actually leads to an improvement. Setting Aims Improvement requires setting aims.  The aim should be time-specific and measurable, with a defined population. Selecting Changes All improvement requires making changes, but not all changes result in improvement. Organizations therefore must identify the changes that are most likely to result in improvement. Testing Changes The Plan-Do-Study-Act (PDSA) cycle is shorthand for testing a change in the real work setting — by planning it, trying it, observing the results, and acting on what is learned. This is the scientific method used for action-oriented learning. A Model for Improvement


Features of Good Aim Statements: Features of Good Aim Statements Specific Measurable Aggressive yet Achievable Relevant Time-bound


Sample Aim Statements:: Sample Aim Statements: Glycemic Control on the Wards Within 6 months the use of sliding scale only regimens will be reduced by half. Within 12 months the % of patients with POC glucose testing achieving a mean glucose of < 200 mg/dL will improve from 65% to 85%. Within 12 months the % of our patients suffering from hypoglycemic events will be reduced from 11% to 6%.


Measurement Principles : Measurement Principles Seek usefulness, not perfection Integrate measurement into daily routine Use qualitative and quantitative data Use sampling Plot data over time Use a balanced set of measures for all improvement efforts


A Blend of Measures: A Blend of Measures Structure Do you have a multidisciplinary steering committee? Do your SQIO sets include a prompt for A1c? Process % of SQIO written using your order form % with basal insulin Outcomes LOS, Mortality: Glycemic control, Hypoglycemia


Picabo Street and Communication: Picabo Street and Communication Olympic Gold Medal Winner….AND a Critical Care Nurse!


Slide19: “Picabo, ICU”


Hierarchy of Reliability: Hierarchy of Reliability No protocol* (“State of Nature”) Decision support exists but not linked to order writing, or prompts within orders but no decision support Protocol well-integrated (into orders at point-of-care) Protocol enhanced (by other QI and high reliability strategies) Oversights identified and addressed in real time Level 4 1 2 3 5 Predicted Success rate 40% 50% 65-85% 90% 95+%


Order sets w/ embedded insulin orders: Standardization?: Order sets w/ embedded insulin orders: Standardization?


High Reliability Design Solutions (as applied to Insulin Protocol): High Reliability Design Solutions (as applied to Insulin Protocol) Standardize insulin choices for common situations MD must “opt out” of default choices (not opt in) Prompts for basal insulin if over glycemic target, prompts for HgA1c, etc. Scheduled assessments of glycemic control / insulin regimen Redundant responsibility to maintain glycemic target


CAUTION!!!! Be Sure to Insert a Brain Between Protocol and Patient!: CAUTION!!!! Be Sure to Insert a Brain Between Protocol and Patient! Education for broad range of providers Consider special team of focused providers


Engineering Change: Hints for Success: Engineering Change: Hints for Success Empower nursing Expedite passage through medical staff committees Better to implement an imperfect, compromise change than no change at all Provide hot line or support for difficult situations Follow metrics continuously as you implement


Engineering Change: Hints for Success: Engineering Change: Hints for Success Measure, learn, and over time eliminate variation arising from professionals; retain variation arising from patients Keep big picture in mind Negotiate ‘speed bumps’ Time delays in getting data Incomplete buy-in Go around obstacles instead of through them (can always go back to them later) Some who disagree with you may be correct Make changes painless as possible: make it easy to do the right thing


PDSA: Plan-Do-Study-Act: PDSA: Plan-Do-Study-Act The use of PDSA has been referred to as the “democratization of the scientific method.” (Paul Miles, MD) Do small scale tests of change. Everyone can do it!


Benefits of rapid cycle change: : Benefits of rapid cycle change: Increases belief that change will result in improvement Allows opportunities for “failures” without impacting performance Provides documentation of improvement Adapts to meet changing environment Evaluates costs and side-effects of the change Minimizes resistance upon implementation


Examples: integration of best practice: Examples: integration of best practice A1c level within last 30 days. Specify hyperglycemic diagnosis Each patient should have a glycemic target.


A1c Level: A1c Level Incorporate prompt for A1c level in insulin order sets and protocols. Ordering can be accomplished with checkbox Monitor performance, feedback to providers Glycemic control team obtains it


Proper diagnosis: Proper diagnosis Diagnosis: Uncontrolled –or–  Controlled Diabetes type:  1  2  Gestational –or– Secondary to another cause;Specify –or–  Stress/situational hyperglycemia Improves reimbursement: define “uncontrolled DM” and monitor coding accuracy Order set docmentation translates into ICD-9


Identify non-critical care glycemic target: Identify non-critical care glycemic target Preprandial target 90–130 mg/dL; maximum random glucose < 180 mg/dL (ADA/AACE consensus target) 80–150 mg/dL Preprandial target 90–130 mg/dL for most patients, 90–150 mg/dL if hypoglycemia risk factors


“Actionable” Glycemic Target: “Actionable” Glycemic Target The “what” is common to all institutions: push for changes in regimens when glycemic target not being met. Variable by institution: Glycemic target definition How to generate report Who acts on report Putting this in place moves you up hierarchy of reliability. Opportunity to Learn from variation!


Hierarchy of Reliability: Hierarchy of Reliability No protocol* (“State of Nature”) Decision support exists but not linked to order writing, or prompts within orders but no decision support Protocol well-integrated (into orders at point-of-care) Protocol enhanced (by other QI and high reliability strategies) Oversights identified and addressed in real time Level 4 1 2 3 5 Predicted Success rate 40% 50% 65-85% 90% 95+% eliminate variation arising from professionals; retain variation arising from patients


Setting: Setting Academic teaching medical centers with over 400 beds Adult inpatients on non-critical care wards with POC glucose testing. Nov 2002 – Dec 2005 Excluded: Critical care, OB, Psych, Senior Behavioral Health


Questions : Questions What is current state? Baseline Nov ’02-Oct ’03. Insulin Use Patterns Glycemic Control Hypoglycemia Other What is effect of implementing a standardized SQIO set? Main Intervention #1 Nov ’03-May ‘05 What is the incremental effect of an insulin management protocol? Main Intervention #2 May ’05-Dec ‘05


Intervention #1 (Nov 2003):A Basic Subcutaneous Insulin Order Set: Intervention #1 (Nov 2003): A Basic Subcutaneous Insulin Order Set Basal / Nutritional / Correction dose terminology introduced Multiple correction dose scales available, based on total insulin dose required. Sliding scale only regimens discouraged Check box simplicity Some guidance for dosing and adjustment Hypoglycemia protocol incorporated Paper, then CPOE versions


Intervention #2 (May 2005)Insulin Management Protocol: Intervention #2 (May 2005) Insulin Management Protocol One page algorithm Glycemic Target Prompt for A1C DC Oral Hypoglycemic Agents Guidance on dosing Suggested regimens for eating patient, NPO patient, patient on enteral nutrition Guidance on dosing adjustment Introduced with case based teaching


The Use of Basal Insulin Increases(sliding scale only regimens decline): The Use of Basal Insulin Increases (sliding scale only regimens decline) 30-90 patients sampled per month, no formal analysis done, results sustained


Glycemic Control: Glycemic Control Days 1 – 14 of admission Exclude patients with < 8 POC tests 5,800 patients 37,516 patient days 111,473 POC tests By patient stay % of patients with mean glucose < 180 mg /dL By patient day % of patient days when all glucose values were between 60 – 180 mg / dL Pearson chi-square statistic to compare: TP 1 (Baseline) Nov ’02 – Oct ‘03 TP2 (Order Set) Nov ‘03 – Apr ’05 TP3 (Algorithm) May ’05 – Dec ’05


Slide41: 62 % 69% 73% 5800 patients w/ > 8 POC glucose values, day 1-14 values: p value < .02 (Pearson chi-square statistic)


Slide42: 1st order set Baseline Order set Algorithm


Slide43: 44% 48% 53% 37,516 Patient Days monitored in 5800 patients with > 8 POC glu tests, day 1-14: (Pearson chi-square statistic p < .001)


Slide44: Clinical Inertia Improves with Order Set and Algorithm


Slide45: Oh no! What about…… HYPOGLYCEMIA!


Hypoglycemia: Hypoglycemia All non critical care patients with POC values 11,057 patients / 53,466 days / 148,466 POC tests Hypoglycemia: ≤ 60 mg/dL Extreme Hypoglycemia: ≤ 40 mg/dL By patient day % of patient days with one or more hypoglycemic events Pearson chi-square statistic to compare: TP 1 (Baseline) Nov ’02 – Oct ‘03 TP2 (Order Set) Nov ‘03 – Apr ’05 TP3 (Algorithm) May ’05 – Dec ’05


Slide47: Percent of Patient Days with Hypoglycemia / Extreme Hypoglycemia decreased by 30% and 31%, respectively. (Pearson chi square p 53,000 patient days > 148,000 POC glu tests


Slide48: Approximately 100 fewer patients with Hypoglycemia per year Month


Summary: Summary Large opportunities for improvement A safety and quality issue Systems approach is needed SHM and others now provide resources to assist implementation teams with all “essential elements” Use “Talking Points”, local anecdote, and small sample data to gain institutional support Reduced hypoglycemia can be compatible with improved glycemic control on the wards Controversy exists, but time for action is now


Slide50: The first time subcutaneous insulin is ordered, the prescriber is asked for an actionable glycemic target. A prompt to order HbA1C is also presented.


Slide51: The weight and markers of insulin sensitivity are elicited, as well as the form of the patient’s nutritional intake. (in this case, the patient is an obese 80 kg woman eating regular meals)


Slide52: The Total Daily Dose (TDD) is calculated for the clinician, based on the information provided on the patient’s obesity and weight. The TDD can be adjusted by the physician. Alternate methods of calculating the TDD are also presented.