JAMIA 20007:186195

1
JAMIA 20007186-195
2
(No Transcript)
3
The Clinical Problem

• The UCLA experience (1992-1995)
• 37,000 pts/yr in ED
• 5,500 pts/yr are children
• 1,400 pts/yr are lt 3 yrs and febrile

4
Some Background

• Fever in children is a common problem
• Children lt3 years are especially difficult
• Cant talk indirect history from parents
• Dont localize well on physical exam (some signs
  absent)
• Immature immune systems
• Under some investigation for the past 20 years
  (Teele, McCarthy started in the 70s)
• Primary bacterial pathogens
• S. pneumonia (pneumonia, meningitis,OM)
• H. influenza (meningitis, epiglottitis, OM)
• E. coli (UTIs)
• N. meningiditis (meningitis, meningococcemia)

5
Enter the 80s and things started to change

• What changed??
• Availability of Ceftriaxone in early 80s
• Long acting 3rd generation cephalosporin
• Replaced ampicillin and chloramphenicol
• H. influenza vaccine became available in 1988
• Primary pathogens
• S. pneumonia (pneumonia, meningitis,OM)
• E. coli (UTIs)
• H. influenza (meningitis, OM) reduced by 90-99
• S. aureus, S. group A, Salmonella sp.

6
The problem of the non-toxic febrile child
without source - 1

• So whats toxic?
• Lethargic
• Poor perfusion (capillary refill)
• Hypo/hyperventilation
• Cyanotic
• So if youre not, but youre febrile 39 C
• Bacteremic 3-11
• S. pneumonia 70
• H. influenza 15
• N. meningitis 3
• Non-bacteremic (viral?) 89

7
The problem of the non-toxic febrile child
without source - 2

• And if youre bacteremic (and not treated)..
• persistent fever - 35
• persistent bacteremia - 12
• meningitis - 7
• UTI (male 6 mo) - 7
• UTI (female 24 mo) - 8

8
Evaluation of the non-toxic febrile child 3 mo
3 yr

• Baraff LJ, et al. Pediatrics 199392,1-12
• Meta-analysis of papers from 1977-91
• General Assessment (Yale criteria)
• CBC 5x risk if WBC 15,000
• Other non-specific test - diff , CRP, ESR ??
• Blood culture
• Lumbar puncture ??
• Urinalysis / culture (Gm stain sediment 99
  sensitive)
• Stool culture diarrhea only
• Chest X-ray temp 40C

9
Treatment of the non-toxic febrile child 3 mo 3
yr

• Baraff LJ, et al. Pediatrics 199392,1-12
• Meta-analysis of papers from 1977-91
• Ceftriaxone 50 mg/kg IM (use 1 lidocaine to
  dilute!)
• Assumes
• IM/IV more efficacious than PO antibiotics
• Dermatologic side effects - 1.5
• GI side effects 4.2
• Anaphylaxis 10-15 in penicillin allergic patients

10
All - CBC
Not obvious whether Option 1 or 2 used
11
EDECS

• Emergency Department Expert Charting System
• 5 modules
• Exposure to body fluids
• Low back pain
• Fever in children lt 3 yrs
• Fundamental concept - if guidelines are to
  useful, they must be integrated in to practice

12
EDECS - 2

• Screen-based - HPI, PMH, PE, etc.
• Color coded essential items in red
• All essential items must be answered
• Unknown fields exist essential items
• Mouse / pull down menus / pick lists
• Lab Data review screens color coded for
  recommendation
• Rationale buttons available
• Advice given for initial/additional testing,
  treatment and follow-up
• Non-compliance explanation requested

13
EDECS - 3

• At visit end
• Prescriptions printed
• Follow-up instructions printed
• Medical record printed
• Selected data stored

14
Previous Experience

• Schriger DL, et al. JAMA 1998270(15) 1339-46
• Occupational exposure to blood and body fluids
• Improved documentation by 42
• Appropriate testing increased by 20
• Appropriate tx decisions increased by 13
• Costs dropped 28

15
Goals of this module

• Improved documentation
• For its own sake
• 2nd most common cause of malpractice suits
• Standardize management of fever without source
• Inclusion of a pediatric module due determine
  unique implementation needs of this age group

16
The Study

• Setting UCLA EM Department
• Off-on-off interrupted time series
• Prospective / retrospective???
• Phase 1 - handwritten - 5/92 12/93
• Phase 2 CPR handwritten 11/94 3/95
• Phase 3 - handwritten 9/95 12/95

17
Elgibility

• Age lt 3 years
• Temp 38C
• Febrile illness
• Exclusions
• Underlying disease (transplant, CHD, renal
  failure, sickle cell disease)
• Localizing respiratory illness (bronchospasm,
  stridor)
• Scheduled recheck

18
Caregivers

• 1992-93 primarily Pediatric Residents (PGY 1-3)
• 1993-95 36 EM residents, occ Peds/FP (PGY 2-4)
• Attendings mainly supervisory
• No attempt to force EDECS use or guideline
  compliance
• 15 minute introduction to EDECS
• Data collection form provided use optional
• Data entered in charting area
• No consent assumed to be normal part of care

19
Encoding and running the guideline

• Guideline encoded using rule-based Application
  Manager
• Intelligent Environments (Tewksbury, MA)
• Proprietary Universal Rules Language to define
  medical logic
• Rule modules
• Check adequacy of documentation, selecting tests,
  choosing diagnoses, preparing follow-up
  instructions
• All modules run with each updated screen

20
Dependent Variables

• Quality of documentation MR follow-up
  instructions
• 21 essential items - what of charts had each
  one
• Overall score of the total number of items /
  chart
• Appropriateness of testing, treatment decisions,
  diagnoses
• Number of appropriate decisions / number of
  decisions
• of testing vs. treatment charges
• Total patient charge
• constant per charge (physician/facility, lab,
  treatment)

21
Data collection, etc

• All CPR charts counted
• Handwritten charts 175 items abstracted
• Phase 1 50 sampled
• Phase 2 100 sampled
• Phase 3 20 sampled
• Sampled to extend study time / reduce cost
• Abstractors trained on std cases, tested for
  inter-rater reliability
• Lab data downloaded
• Case complexity established using age, triage
  note, first line of chart note

22
Data collection, etc

• Expected 50 compliance
• Designed to have 90 power to detect an increase
  to 80
• Patient as unit of analysis as most caregivers
  were present in only one phase of the study
• Categoric variables (s) logit procdure
• Between-phase differences - 95 CIs
• Linear regression for charges (raw
  log-transformed)

23
Results
Charts evaluated 830/1630
Physicians involved 185 1 phase - 141 2
phases - 35 3 phases - 9
Physician / Patient Load 50 saw 1-2
patients Average load 4.5 patients Range - 1-32
patients
24
(No Transcript)
25
76 Physicians Involved in Phase 2 32 never
used EDECS 25 who saw more than 5 patients
accounted for 64 of use From previous table, EM
residents more likely to use (68)
26
Phase 2 includes hand-written and EDECS
27
Improvement in PE - not significant

• Pneumotoscopy
• Phase 1 15
• Phase 2 (handwritten) 10
• Phase 2 (EDECS) 20
• Phase 3 5

28
Phase 2 includes hand-written and EDECS
29
Required 2/4 EDECS criteria to be met Overridden
twice
30
Appropriateness correct decisions

• Diagnostic Tests
• CBC 84
• Blood culture 83
• UA 81
• Urine culture 84
• Chest X-ray 70
• Lumbar puncture 86
• Oral Antibiotics (give or withhold) 96
• 58 of patients discharged home on these
• Ceftriaxone 44 evidence of FWS
• 10 of patients discharged home after IM dose
• No difference in charges or charge mix (adjusted)

31
Discussion

• Documentation improved
• Evaluation of otitis media improved
• Temporal correlation with improved diagnosis?
• Residents farther along in training?
• Awareness that viral syndrome is common
• Guideline compliance unchanged

32
Why Physicians Dont Follow Guidelines

• Cabana, et al. JAMA 19992821458-65
• Lack of awareness
• Lack of familiarity
• External barriers (guideline, patient,
  environmental)
• Inertia of previous practice
• Lack of self-efficacy
• Lack of Agreement

33
The Controversy why the guideline may not be
followed -1

• Kramer MS Shapiro ED. Pediatrics
  1997100128-133
• Risk/benefit decision Provider vs. Parent
• Why one wouldnt use the guideline
• Low PPV of WBC for bacteremia 8-15, even less
  for serious infections (meningitis,
  osteomyelitis)
• Blood culture may not predict meningitis, so why
  use it?
• Many infants will clear their bacteremia
  spontaneously resulting in an unnecessary
  hospitalization
• Bacteremia due to H. influena reduced since
  immunization
• Trauma to child and parents
• Primary follow-up is better than that of the ED

34
The Controversy why the guideline may not get
followed -2

• Whittler RR, et al. Pediatr Infect Dis J.
  199817271-7

35
The Controversy why the guideline may get
followed - 1

• Baraff LJ, et al. Pediatrics 1997100134-136
• Risk/benefit decision Provider vs. Parent
• Why one would use the guideline
• Yes, all the above are true, but
• Giving antibiotics to S. pneumonia bacteremic
  children results in
• Less febrile infants (27 vs. 78)
• Less likely to be admitted (14 vs. 50)
• Potentially less morbidity/mortality

36
The Controversy why the guideline may get
followed 2

• Baraff LJ, et al. Pediatrics 1997100134-136
  (cont)
• Is it worth subjecting
• 600 children to a CBC
• 200 children with WBC 15,000 to Blood Culture
  and Ceftriaxone
• To prevent 1 case of meningitis (your call)
• Clinically consider the physicians biases and
  beliefs, as well as the parents
• Fleisher Bass 7525 children (all treated)
• Est. 215 cases of S. Pneumonia bacteremia present
• Est. 11 cases S. Pneumonia meningitis prevented

37
So what have we learned?

• Pick your guideline carefully (dont expect rigid
  compliance, after all it is a guideline)
• Re-evaluate your guideline periodically (what
  will happen with the new pneumococcal vaccine?)
• Dont miss then opportunity to learn something
  (the opt-out data may have been here, but wasnt
  used)
• If at all possible, randomize (by provider,
  patient, practice)
• Avoid confounding variables (time was critical
  here several reasons)
• Dont ignore the cost of provider time (what was
  the cost of improved documentation if no
  compliance change occurred)