clinical practice: the only constant is change practice: the only constant is change brent c. james,...
TRANSCRIPT
Clinical Practice:The Only Constant Is Change
Brent C. James, M.D., M.Stat.Executive Director, Institute for Health Care Delivery ResearchIntermountain Health CareSalt Lake City, Utah, USA
IHC ATP Alumni Conference16th Annual IHI National Forum on Quality Improvement in Health Care
World Center Marriott Resort & Convention Center, Orlando, FloridaMonday, 13 December 2004, 6:00p - 9:00p
Dr. John Wennberg
Geography is destiny ("Who you see is what you get" *)
There is no health care "system"Supplier-induced demand:
Field of Dreams approach: Build it and they will comeJames T. Kirk: Do something, Bones! She's dying!Eddy: More is better -- if it might work, do itChassin: Enthusiasm for unproven methodsBoston City / Boston University Hospital, 1998:
Same housestaff on both servicesMore beds / easier access to resources on Boston University serviceBoston University readmit rate ~50% higher
* Richard Deyo, MD, MPH - in: Cherken, Deyo, Wheeler and Ciol. Physician variation in diagnostic testing for low back pain. Arth & Rheum 1994; 37(1):15-22 (Jan).
American health care"gets it right"
54.9%of the time.
McGlynn EA, Asch SM, Adams J, et al. The quality of health care delivered to adults in the United States. N Engl J Med 2003; 348(26):2635-45 (June 26).
Medical injuries
Account for44,000 - 98,000 deaths per year
in the United StatesMore people die from medical injuries than from
breast cancer or AIDS or motor vehicle accidentsBrennan et al. New Engl J Med 1991 Thomas et al. 1999
Current health care
is the best the world has ever seenA few simple examples:
— From 1900 to 2000, average life expectancy at birthincreased from only 49 years to almost 80 years.
— Since 1960, age-adjusted mortality from heart disease (#1) has decreased by 56%; and (from 307.4 to 134.6 deaths / 100,000)
— Since 1950, age-adjusted mortality from stroke (#3) has decreased by 70%. (from 88.8 to 26.5 deaths / 100,000)
.
Initial life expectancy gains almost all resulted from public health initiatives -- clean water, safe food, and (especially) widespreadcontrol of epidemic infectious disease. But since about 1960, direct disease treatment has make increasingly large contributions.
Centers for Disease Control. Decline in deaths from heart disease and stroke--United States, 1900-1999. JAMA 1999; 282(8):724-6 (Aug 25).National Center for Health Statistics. Health, United States, 2000 with Adolescent Health Chartbook. Hyattsville, MD: U.S. Dept. of Health and Human Services, Center for Disease Control and Prevention, 2000; pg. 7 (DHHS Publication No. (PHS) 2000-1232-1).U.S. Department of Health and Human Services, Public Health Service. Healthy People 2000: National Health Promotion and Disease Prevention Objectives. Washington, DC: U.S. Government Printing Office, 1991 (DHHS Publication No. (PHS) 91-50212).
Medicine used to be simple, ineffective, and relatively safe.
Now it is complex, effective, and potentially dangerous.
Neal G. Reducing risks in the practice of hospital general medicine. In Clinical Risk Management, 2nd edition. British Medical Journal, 2001.
Chantler, Cyril. The role and education of doctors in the delivery of health care. Lancet 1999; 353:1178-81.
Sir Cyril Chantler
Are most failures unavoidable?
The price we pay(for)
diseases of medical progress
Barr, David. Hazards of modern diagnosis and therapy - the price we pay. JAMA 1955; 159(115):1452-6 (Dec 10)Moser, Robert H. Diseases of medical progress. N Engl J Med 1956; 255(13):606-14 (Sep 27)
Reasons for practice variation
ComplexityHow many factors can the human mind simultaneously
balance to optimize an outcome?"The complexity of modern American medicine exceeds
the capacity of the unaided human mind"
Lack of valid clinical knowledge (poor evidence)
-- Alan Morris, MD
-- David Eddy, MD
Reliance on subjective judgmentSubjective evaluation is notoriously poor across groups
or over time
Mark Chassin, MDEnthusiam for unproven methods ...If it might work, do it ...Quality = spare no expense ...
David Eddy, MD, PhD
Brent James, MD, MStat
Clinical uncertainty:
The core assumption
"Our minds are interpreters of evidence. We can accurately convert all forms of evidence (formal evidence, observations, experiences, colleague's experiences) into conclusions, which in turn determine our actions."
"Therefore, no one has to tell us what to do. Just give us the evidence and we will figure it out. Besides, there are lots of other factors that need to be considered. This can only be done with clinical judgment."
Evidence Ourminds Conclusions Actions
Dr. David Eddy
The core assumption is untenable
Poor evidence
The inherent complexity of modern medicine, versus the limitations of the human mind
Variations in beliefs
Variations in practices
High rates of inappropriate care
Dr. David Eddy
Other factors affect our decisions
Evidence Ourminds Conclusions Actions
Dr. David Eddy
Professional interestsFinancial interestsPersonal tastesDesire to have something to offerLove for the workWishful thinkingSelective memoryPressure from patients and familyLegal considerations
If our minds can't do the work very well, there are all sorts of other things to fill the void:
Huge ranges of uncertaintyLimited, complex Massive variation,inappropriate care
Protocols can improve care
A multidisciplinary team of health professionals -1. Select a high priority care process2. Generate an evidence-based "best practice" guideline3. Blend the guideline into the flow of clinical work
staffingtrainingsuppliesphysical layoutmeasurement / information floweducational materials
4. Use the guideline as a shared baseline, with clinicians free to vary based on individual patient needs
5. Measure, learn from, and (over time) eliminate variation arising from professionals; retain variation arising from patients ("mass customization")
From craft-based practiceindividual physicians, working alonehandcraft a customized solution for each patientbased on a core ethical commitment to the patient andvast personal knowledge gained from training and
experience
To profession-based practicegroups of peers, treating similar patients in a shared settingplan coordinated care delivery processeswhich individual clinicians adapt to specific patient needs
(e.g., standing order sets)
(housestaff ::= apprentices)
early experience shows less expensiveless complexbetter patient outcomes
(facility can staff, train, supply an organize to a single core process)(which means fewer mistakes and dropped handoffs, less conflict)
The healing professions - and health care delivery -are changing ...
Idea 1
The solo practice of medicine is dead(at an intellectual, not necessarily an economic, level)
protocols as a common baselinemass customizationhigher productivity (a way to defend income)better patient outcomes (closing the quality gap)strong ties to electronic medical records
Issues for data automation
Principle: Automation must improve productivity;you can't destroy clinical work flow!
Issue 1: Paper vs. electronic storage
Issue 2: Free-text vs. encoded data
Four types of information
1. Sound data (e.g., recorded voice)
2. Image data (pictures, stored as rasterized pixel bitmaps; e.g.,a picture of a feature on a patient's skin;a digitized AP chest x-ray; or aa picture of a handwritten or typed medical note.)
4. Encoded data (multi-field "concepts" that computers can compare, interpret, etc.; required to do clinical decision support)
3. Free-text data standard word processor output;characters stored as codes, one per byte;even numbers are stored as characters;using ASCII codes (American Standard Code for
Information Interchange)
natural language
processing
optical character
recognition
computerized voice recognition
Clinical Data Repository (CDR)
relational data base management system
(rdbms)
EMPI -electronic master
patient index
VocSer -vocabulary server
PCMS(inpatient EMR)
Results Review Clinical
Workstation(outpatient EMR)
Pathology - laboratory - microbiology -surgical pathology
Pharmacy(operations)
Imaging - reports - images
Financial transactions
Inventory relief
EDW link - patient registries - outcomes reporting - etc.
Levels of data automation
2. Automation of lab - (encoded) clin path - microbiology - surgical path pharmacy - 3 national services, standard (encoded) formats
consultation reports - HL-7 formatted text
imaging - HL-7 formatted text (active research --> encoding)- digital image storage
3. The 'electronic file cabinet' - all electronic record (don't care about encoding)efficient data entry - dictation/transcription --> 'hot text,' voice recognitionefficient storage and retrieval across many care sitespractice operations support - telephone call support, lab results management
5. Level 2 encoding - focused condition-specific encoding
6. Level 3 encoding - fully encoded EMR, with text comments to customize
4.encoded problem list - automatic 'registries' for common chronic diseasesencoded medication list - first easy step in computerized order entryencoded allergy list
Level 1 encoding (assumes encoded patient demographics, pharmacy, and lab)
1.scheduling
Automation of billing encoded patientdemographics
message logsInfoButtonResults Reviewlab management
Implementing CW hot textAverage WRVUs per Day Average Patient Wait Time
Dr. Mark Shepherd, Bryner Clinic
21.323
24.1 24.323.2
24.523 23.6 23.4 23.7
31.2
21.9
16.7 16.9 16.915
16.9 16.9
14.7
17.2
QI, Q2 2
000
Q3 200
0Q4 2
000
Q1 200
1Q2 2
001
Q3 200
1Q4 2
001
Q1 200
2Apr 2
002
May 20
02
0
5
10
15
20
25
30
35
Ave
rage
Pat
ient
Wai
t Tim
e (m
inut
es)
0
5
10
15
20
25
30
35
Ave
rage
WR
VUs
per D
ay
Implementing CW hot text
Dr. Mark Shepherd, Bryner Clinic
293
369
299
349 337
371
335317
378
331 319342 343
420
358385
13531306
444402 405
286207
11021 10 0 0 1 0 0 0
Sep 99
Oct 99
Nov 99
Dec 99
Jan 00
Feb 00
Mar 00
Apr 00
May 00
Jun 00
Jul 0
0Aug 00Sep
00Oct
00Nov 0
0Dec
00
0
300
600
900
1200
1500
Tota
l Tra
nscr
iptio
n C
ost p
er M
onth
0
100
200
300
400
500
Tota
l WR
VUs
per M
onth
Total Transcription Cost per Month Total WRVUs per Month
Idea 2
EMRs are essential tools for the future(1) productivity enhancement; (2) protocol implementation
hard to implement (at present time)requires a relatively large organizationwe are in the era of the big systemsnational effort to (1) establish standards (HIPAA was just the beginning) and(2) drive EMR adoption (Regional Health Information Infrastructure)
Community acquired pneumonia
protocol protocol
"Outlier" (complication) DRG at discharge
15.3% 11.6%
In-hospital mortality
7.2% 5.3%
Relative resource units (RRUs) per case 55.9 49.0
Cost per case $5211 $4729
without with
24.7%
26.3%
12.3%
9.3%
p<0.001
p=0.015
p<0.001
p=0.002
Impact on net income
Improvement tocost structure
Payment mechanism
Decrease cost per unit
Decrease LOS (# nursing hours)
Decrease # of cases
Decrease # units per caseDecrease other units per
case
DiscountedFFS
(45%)
Per case
(40%)
Shared risk
(15%)
Per diem
(0%)
Impact on net income
Improvement tocost structure
Payment mechanism
Decrease cost per unit
Decrease LOS (# nursing hours)
Decrease # of cases
Decrease # units per caseDecrease other units per
case
DiscountedFFS
(45%)
Per case
(40%)
Shared risk
(15%)
Per diem
(0%)
Idea 3
Pay for performance (P4P):
one form of outcomes accountabilityrelies of performance measurement (technical problems)two main forms:- hit performance goal, get %-age payment increase- shared savings models
High frequency injuries sources
1. Adverse drug events (ADEs, ADRs)
post-operative deep wound infectionsurinary tract infections (UTI)lower respiratory infections (pneumonia or bronchitis)bacteremias and septicemias
2. Iatrogenic infections
4. Mechanical device failures3. Pressure injuries
5. Complications of central and peripheral venous lines
6. Deep venous thrombosis (DVT) / pulmonary embolism (PE)
7. Strength, agility and cognition (injuries and restraints)
8. Blood product transfusion
9. Patient transitions
Detecting Adverse Drug Events
# of ADEs / %(# per
annum)
Total ADEs
Nurse Incidence Reporting
"Enhanced" Reporting
HELP System
Moderate and severe
ADEs
9 / 0.025%(6)
91 / 0.25%(60)
731 / 2.0%(487)
701 / 1.9%(467)
Simple criteria for detecting ADEs
use of naloxoneuse of benadryluse of inapsineuse of lomotilnurse reports of rash/itchinguse of loperamidetest for c. deficile toxindigoxin level > 2abrupt med stop or reductionuse of vitamin Kdoubling of blood creatinineuse of kaopectateuse of paregoricuse of flumazenil
pharmacypharmacypharmacypharmacynurse reportingpharmacyclinical labclinical labpharmacypharmacyclinical labpharmacypharmacypharmacy
21.9 28.3 28.321.0 20.8 49.139.2 20.4 69.526.8 8.5 77.017.9 5.1 82.122.3 3.4 85.524.3 3.1 88.6 2.3 2.2 90.848.0 1.0 91.8 4.8 0.9 92.7 0.4 0.8 93.521.8 0.7 94.2 9.8 0.7 95.077.3 0.7 95.7
1.2.3.4.5.6.7.8.9.
10.11.12.13.14.
Detection criterion Location Rate (%)Positive
True
DetectedAll ADEs
% of
Total (%)Cumulative
(case finding through "concurrent clinical triggers")
Utah-Missouri study (AHRQ)
Identified and validated ICD-9 Patient Safety Code set (~1,000 codes)
Mined hospital discharge abstracts to identify charts for review (verified acceptable positive predictive value)
Found numbers of verified injuries similar to those discovered by concurrent trigger tools (including serious outpatient events that result in hospitalization)
Very little overlap with concurrent trigger methods: ICD-9 codes appear to capture a different universe of events
(case finding through "retrospective code triggers")
LDS Hospital, 2001
ICD9-CM total code hits = 1574(973 unique)
concurrent trigger total hits = ~2460(~1950 unique)
(major advantage: can intervene and abort harm)
pharmacist verified ADEs from concurrent triggers = 258
(168 unique)
511unique
90unique
200 charts in random sample- 13 charts couldn't be found (6.5%) 187 charts reviewed
10 coding errors ( 5%) 24 not ADEs (false positives) (13%) 51 outpatient poisonings and suicides (27%) 59 outpatient ADEs (32%) 43 inpatient ADEs (true positives) (23%) (1484 x 0.230 = 341 unique inpatient ADEs)
Final estimates: 341 ADEs detected by ICD-9 codes alone 90 ADEs detected by both systems 168 ADEs detected by concurrent triggers alone
599 total ADEs detected (43.1% of final total detected by concurrent triggers alone)
Xu W, Hougland P, Pickard S, et al. Detecting adverse drug events using ICD-9-CM codes. Poster presentation, AHRQ 2nd Annual Patient Safety Research Conference, Arlington, Virginia, 3 March 2003.
random sampleof 200 encounters(from 1484 cases)
Tracking injuries
Current (voluntary reporting) systems miss the vast majority of injuries (finding only 1 in 100-150 injuries)
Most often (e.g., >80% of the time for ADEs), clinical teams don’t associate patient symptoms with the treatments that are causing them
A more accurate perception of sources of injury can hugely change intervention strategies
Injury and near miss detection
- case finding - evaluation - classification
Epidemiologic analyses
- new taxonomies
Process Analyses - FMEA - HACCP - other related methods
Successful change ideas from the medical literature, other health care organizations, and other industries
Possible system fixes that could improve safety results
- hypotheses for process change - explicit process models very useful
Prioritization (is this event an epidemic?)
Rapid Cycle Testing often with heavy reliance on
intermediate outcomes (process measures)
Holding the Gains Monitor final outcomes (i.e., injury rates) and
intermediate process outcomes
Deployment and Implementation
Patient Safety: Achieving a New Standard for Care. Edited by Aspden P, Corrigan JM, Wolcott J, and Erickson SM. Washington, DC: National Academy Press, 2004. Figure 5, pg. 179.
Detecting patient safety events
1. Case finding (based on explicit criteria)
2. Evaluation (based on explicit criteria)
3. Classification (based on explicit criteria)
Externaldata system
audit
Case finding
Common eventsAdverse drug eventsIatrogenic infectionsPressure injuryMechanical device failuresVenous linesVTETransfusionsPatient fallsPatient transitions
Rare events "Wrong" surgeryKidnappingSuicide - etc. -
Concurrent(data-based)
trigger systemsLDSH / B&W
CDC infection controlx
LDSHxxx
B&Wx
Abstract (ICD-9)(data-based)
trigger systemsUtah-Missouri (AHRQ)
xx
LDSHxxxxx
Criteria-basedmanual systems
(e.g, QaRNS, JCAHO SE,NQF "Never Events")
Relatively poor event detection
rates
Current most
effective method
Voluntaryreporting(in a Culture
of Safety)
Relatively poor event detection
rates
Current most
effective method
Retrospective Chart Review
Preventable causes of ADEs
PhysiologicFactors
PharmocologicFactors
Drug Administration
Errors
Ordering Errors
Transcribing
Spelling
HemalRenal
Dilution
Time
Nurse
Route
Rate
ADE
NursePhysician
Pharmacist
PhysicianPharmacy
Nurse/Clerk
PharmacistPatient
PhysicianDietician
Physician
Wrong Drug
Dose
Scheduling
Dosage
Route
Past Allergic Reaction
Absorption
WeightAge
Gender
Electrolyte
Hepatic
RaceBrand Name vs. Generic
Drugs
Expected
Drug/Drug
Unforeseen
Drug/Food
Drug/Lab
Neural
Psychic
Compliance
Patient
OrderMissed
Causes of Adverse Drug Events
0
10
20
30
40
50
60
70
80
90
100
Per c
ent
Causes
Causes of Adverse Drug Events
Class % Description Avoidable?
?28.0PharmExpected
Known drug reactions
Yes23.0 Failure to adjust for decreased renalfunction
PhysioRenal
Yes14.2 Failure to adjust for patient agePhysioAge
Yes5.7 Failure to adjust for patient body massPhysioWeight
Yes5.0 Error in dosage on orderOrderDosage
Yes4.6 Failure to adjust for known hematologicfactors
PhysioHemal
Totalpreventable 66.2
Medication errors vs. ADEs
Prospective daily surveillance of 202,222 inpatients for theoccurrence of medication errors and adverse drug events
Medicationerrors
n = 4155
Adverse drugevents
n = 3996
n = 138
Definition of medication errors: Assumes that the physician orders correctly,but that the pharmacist then prepares the medication incorrectly, or that the
nurse delivers it incorrectly. Specifically, (1) wrong preparation, (2) wrong dose,(3) wrong route of delivery, (4) wrong rate of delivery, and/or (5) wrong patient.
Attack injuries, not errors
What is classified as an "error "derives from what is judged to be "preventable; "
but, at this stage, those (subjective) judgmentsmay be perverted by the "name, shame, and blame game," and seriously misinformed .
It is more useful to think in terms of"medical injuries" rather than "errors."
ADEs at LDS Hospital
233
581 567 569 567
437
477
355
271 271 280
89 90 91 92 93 94 95 96 97 98 99(3)
Year
0
100
200
300
400
500
600
0
100
200
300
400
500
600
Tota
l # m
oder
ate
+ se
vere
AD
Es
Prophylactic antibiotics on time
10.14 9.8910.19
8.92 8.75
7.867.53
8.83 8.748.11
3.19
7.18
3.2
5.44
4.64
Q1 199
9 Q2 Q3 Q4Q1 2
000 Q2 Q3 Q4
Q1 200
1 Q2 Q3 Q4Q1 2
002 Q2 Q3
0
2
4
6
8
10
12
Org
an s
pace
& d
eep
infe
ctio
ns p
er 1
000
0
2
4
6
8
10
12
n = 9767 9519 9869 9713 9799 10093 9743 9951 10117 9407 9612 93779104 9737 9700
8.97
5.33
Idea 4
An expanded nationalfocus on patient safety
standardized data systemsindependent, external auditdetected injury rates much higher than todaydrive for substantial systems-level change
ICU admissions by weeks gestation
6.66
3.36
2.47 2.65
3.44
4.26
37 38 39 40 41 42
Weeks gestation
0
2
4
6
8
10
Perc
ent N
ICU
adm
issi
ons
0
2
4
6
8
10
Deliveries w/o Complications, 2002 - 2003
8,001 18,988 33,185 19,601 4,505 258n =
RDS by weeks gestation
1.19
0.47
0.25 0.3
0.470.39
1.92
0.68
0.42 0.41
0.670.78
37 38 39 40 41 42
Weeks gestation
0
0.5
1
1.5
2
2.5
Perc
ent
0
0.5
1
1.5
2
2.5
Deliveries w/o Complications, 2002 - 2003
8,001 18,988 33,185 19,601 4,505 258n =
Elective inductions <39 weeks
5.5 5.1
6.6 6.3 65.3
8.2
5.4 5.76.6 6.6
7.9
6.4
7.6 7.6
4.63.5
4.55
26.726.9
29 29.2
25.3
27.6
20.4
19.1
16.5
15.2
8.4
10.7
8.1
6.85.9 6.1 6
5.1
6.3
Jan 01 Mar May Ju
lSep Nov
Jan 02 Mar May Ju
l
Jan 03 Mar May Ju
lSep Nov
Jan 04 Mar May Ju
l
0
5
10
15
20
25
30
% e
lect
ive
indu
ctio
ns <
39
wee
ks
0
5
10
15
20
25
30
382372
490415
430435
422455
430382
356337
372366
455n = 423453
473476 512
475602
557667
564637
578541
573533
505501
474536
562545
535500
Unplanned c-section rates
3331.4
36.1
28.3
17.7
15.1
17.6
14.4 14.3
5.84.5
2.1
0
20
8.2 8.5
3.6 3.4 3.93.2
2.41.1 0.9 1
0 0
1 2 3 4 5 6 7 8 9 10 11 12 13
Bishop score
0
5
10
15
20
25
30
35
40
Perc
ent c
-sec
tions
0
5
10
15
20
25
30
35
40
Electively induced patients by Bishop score, Jan 2002 - Aug 2003
10 49 130 274 567 856 1114 1266 1062 737 415 86 1918 35 61 99 164 278 375 487 453 346 179 47 7
MultipsPrimips
n
Average hours in labor & delivery
22.1
20.7
17.4
15.715
13.8
12.611.6
10.4
9 9
7.58.2
12.4 12
10.810.1
9.2
8.17.6
7.16.4
5.9 5.5 5.14.1
1 2 3 4 5 6 7 8 9 10 11 12 13
Bishop score
0
5
10
15
20
25
Hou
rs
0
5
10
15
20
25
Electively induced patients by Bishop score, Jan 2002 - Aug 2003
10 49 130 274 567 856 1114 1266 1062 737 415 86 1918 35 61 99 164 278 375 487 453 346 179 47 7
MultipsPrimips
n
Primiparous elective inductions
15.314
15.3 14.5 14.7
11.612.8
11.8 12.6 12.8
15.1
12.1
9.98.8
6.8 6.5 6 6.1
53 53
63
5357
45
5652
41
52
62
4649
35
21 2126 28
110
87
119
109
124
91
107
94100
105
118
8781
67
57 57
4652
Jan 20
03 Feb Mar AprMay Ju
n Jul
Aug
Sep OctNov
DecJa
n 2004 Feb Mar Apr
May Jun
0
20
40
60
80
100
120
140
Num
ber o
f pat
ient
s
0
10
20
30
40
50
% o
f all
prim
ipar
ous
deliv
erie
s
Bishop's score < 10Bishop's score < 8Goal: Reduce "inappropriate" nullip inductions by 50%
Labor & delivery variable cost
Jan 20
03 Feb Mar Apr
May Jun Jul
Aug
Sep Oct
Nov
DecJa
n 2004 Feb Mar Apr
May
1000
1200
1400
1600
1800
2000
Ave
rage
com
bine
d va
riabl
e co
st ($
)
1000
1200
1400
1600
1800
2000
Expected maternal and fetal combined variable costGoal: hold increase to no more than 6.85%Actual combined variable cost
Well newborn bilirubin testing
Mar 20
01 May Jul
Sep NovJa
n 2002 Mar May Ju
lSep Nov
Jan 20
03 Mar May Jul
Sep NovJa
n 2004 Mar May
0
20
40
60
80
100
% te
sted
0
20
40
60
80
100
Newborns >= 35 weeks gestation seen in Well Newborn Nursery (excluding hospitals using Bilicheck testing)
0
5
10
15
20
25
0 12 24 36 48 60 72 84 96 108 120
Hour-Specific Bilirubin Risk Chart for Term & Near-Term InfantsAdapted and revised [April 2003] based on IHC data (12-54h) & from Bhutani VK et al. Pediatr 1999; 103:6-14 & J Perinat 2001; 21:S76-S82 (72-120h)
Neo
nata
l Ser
um B
iliru
bin
(mg/
dL)
Age (h)
NSB in 48h†
High Risk Zone
Low Risk Zone
Low Intermediate Risk Zone
High Intermediate Risk ZoneConsider Phototherapy if Premature or
Evidence of hemolysis
95th percentile
75th percentile
40th percentile
NSB in 24h†
Phototherapy & NSB in 6-12h†
NSB >25: Neonatology Phone Consult; Consider Exchange Transfusion in the Healthy Term InfantNSB >20: Consider Exchange Transfusion in the Hemolytic Term Infant or Healthy Near-Term Infant
Risk FactorsJaundice in the first 24hVisible jaundice before dischargePrevious jaundiced siblingGestation < 38 weeksExclusive breastfeedingEast Asian raceBruising, cephalohematomaMaternal age > 25yMale sex
† A TcB may be substituted for NSB. Near exchange levels, a NSB is preferred. NSB = Neonatal serum bilirubin; TcB = transcutaneous bilirubin
Newborns w/ hyperbilirubinemia
01
20
3 3
0
3
02 2
3
02
12 2
12
1 1 12
01
02
13
0 0 01
01 1
0 0 0 0
2826
27
37
26
32
24
34
30
16
34
19
28
2224
2627
3234
31
25
1716
14
27
20
1415
1315
10
1315
12
1615
10
21
13
16
Mar 20
01 May Jul
Sep NovJa
n 2002 Mar May Ju
lSep Nov
Jan 20
03 Mar May Jul
Sep NovJa
n 2004 Mar May
0
10
20
30
40
50
Num
ber o
f pat
ient
s
0
10
20
30
40
50
Bilirubin > 19.9 mg/dLBilirubin > 25 mg/dL
Hyperbilirubinemia readmissions
0.061
0.039
0.034
0.073
0.0470.044
0.027
0.036
0.022
0.041
0.035
0.071
0.044
0.081
0.043
0.027
0.036
0.0450.048
0.062
0.0350.036
0.049
0.019
0.0330.0290.029
0.034
0.022
0.045
0.0220.0230.02
0.044
0.024
0.012
0.018
0.025
0.017
0.009
0.014
0.0090.009
0.02
0.0140.01
0.0260.0220.023
0.021
0.008
0.019
0.008
0.0160.014
Jan 0
0
Apr Jul
OctJa
n 01
Apr Jul
OctJa
n 02
Apr Jul
OctJa
n 03
Apr Jul
OctJa
n 04
Apr Jul
0
0.02
0.04
0.06
0.08
0.1
Prop
ortio
n re
adm
itted
0
0.02
0.04
0.06
0.08
0.1
"I am sorry for you, young men (and women) of this generation. You will do great things. You will have great victories, and standing on our shoulders, you will see far, but you can never have our sensations. To have lived through a revolution, to have seen a new birth of science, a new dispensation of health, reorganized medical schools, remodeled hospitals, a new outlook for humanity, is not given to every generation."
At the opening of the Phipps Clinic in England, near the end of his career. Cited in
-- Sir William Osler
Reid, Edith Gittings. The Great Physician: A Life of Sir William Osler. New York, NY: Oxford University Press, 1931 (p. 241).