Evidence Based Medicine and Level 1 Evidence Based Medicine and Level 1 Outcomes Research in Pediatric Outcomes Research in Pediatric
SurgerySurgery
George W. Holcomb, III, M.D., MBAChildren’s Mercy Hospital
Kansas City, Missouri
Center for Prospective Clinical Trialswww.cmhclinicaltrials.com
• Integration of best research evidence with clinical expertise and patient values
• Treating patients based on data, not “feeling” (gestalt), or one’s own experience
Evidence Based Medicine
Levels of EvidenceQuestion
A clinical surgeon publishes a retrospective review of 350 patients over 20 years undergoing an endorectal pull-through (Soave procedure) for Hirschsprung’s Disease. This is felt to be a seminal paper on this disease in infants and children. What is the level of evidence for this paper?
Level 1 (A)Level 2 (B)Level 3 (C)Level 4 (D)Level 5 (E)
5 – Expert opinion, or applied principles from physiology, basic science, or other conditions
4 – Case series or poor quality case control and cohort studies
3 – Case control studies
2 – Review of case control or cohort studies with agreement or poor quality randomized trial
1 – Prospective, randomized controlled trials
Levels Of Evidence
Levels of Evidence
5. Expert opinion, or applied principles from physiology, basic science, or other conditions
Example:
• Leave patient intubated and paralyzed for 3-5 days following an esophageal resection to take tension off esophageal anastomosis
No data – no study to show intubated/paralyzed patient has less esophageal tension
• Transverse incision is used for abdominal exploration in baby b/c better exposure
No data to support this practice
Levels of Evidence
4. Case series or poor quality case control and cohort studies
Example:
1) Paper reviewing results from one approach to a disease
Large retrospective review of Soave operation for Hirschsprung’s Disease
Levels of Evidence
3. Case control studies
Example:
1) Paper showing different management strategies/operative technique for one disease process
Single center (or multicenter) retrospective review of Duhamel vs Soave operation for Hirschsprung’s Disease
Levels of Evidence
2. Review of case control or cohort (followed “long-term”) studies with agreement or a poorly performed prospective randomized trial
Example:
1) Review of two or three large series describing one management strategy (Soave procedure) compared to two or three large series describing another management strategy (Duhamel procedure)
p - Value• Commonly accepted p-value is 0.05
• 5/100 (1/20) chance that if a difference is found b/w variables, there really is no difference b/w the variables
• If p=0.01 there is a 1/100 chance that if a difference is found b/w variables, there really is no difference
• P=0.07 7/100 – still pretty significant, but readers will think there really is no significance b/c >0.05 (depends on what variable one is looking at)
How To Set Up a Study
• Alpha –Usually 0.05 is used; establishes level of significance of study
Value is used to determine # of patients needed
Retrospective fundoplication review: 12% vs 5%• 360 patients needed to show a significant difference• If αα was 0.1, need fewer patients• If αα was 0.01, need more patients (because significance is
greater)
Set the αα low to avoid a Type I error
How To Set Up a Study
• Power – commonly accepted is 0.8
• 20% chance that the study will fail to detect a difference when there really is one
or
• 80% chance that the study will detect a difference when one does exist
• Underpowering a study risks a Type II error
Children’s Mercy Hospital
Focus on common conditions which are “controversial”:
• Pyloric stenosis
• Appendicitis
• Pectus excavatum
• Fundoplication for reflux
• Empyema
• Non-palpable intra-abdominal testis
Remember
• There is a lot more to an MIS operation than just technique
• Postoperative care is also important and open for study (antibiotics?, pain management?, etc.)
Landscape of Pediatric Surgery
• Young field, most practicing surgeons are 2-3 generations removed from the founders of the field
• Very few training centers, the opinions of a few affect the masses
• Wide variety of rare cases make acquisition of objective treatment data difficult
Open vs Lap Pyloromyotomy
• Lap vs Open – 2003 - controversial around the world and in our hospital
• Different feeding regimens used in our hospital (2 hours, 4 hours, 6 hours)
• Different postoperative pain management strategies utilized
• Differences between staff made it difficult for residents, NPs, floor nurses
• Benefits: single protocol for feeding, pain management, discharge used in study still used currently (6 years later)
• No level 1 data
ResultsOutcomesOutcomes
OPEN (n = 100)OPEN (n = 100) LAP (n = 100)LAP (n = 100) P ValueP Value
OR time OR time (mins)(mins) 19:28 +/- 0.60 19:28 +/- 0.60 19:34 +/- 0.7819:34 +/- 0.78
Emesis (#)Emesis (#) 2.61 +/- 0.27 1.84 +/- 0.23 2.61 +/- 0.27 1.84 +/- 0.23 Full Feeds (hrs)Full Feeds (hrs) 21:01 +/- 2.16 21:01 +/- 2.16 19:30 +/- 1.46 19:30 +/- 1.46
LOS (hrs)LOS (hrs) 33:10 +/- 1.63 33:10 +/- 1.63 29:38 +/- 1.69 29:38 +/- 1.69
Tylenol (doses)Tylenol (doses) 2.23 +/- 0.18 2.23 +/- 0.18 1.59 +/- 0.161.59 +/- 0.16
(Mean +/- S.E.) (Mean +/- S.E.)
0.93
0.05
0.43
0.12
0.01
Ann Surg 244:363-370, 2006 Ann Surg 244:363-370, 2006
ConclusionsConclusions
• Operative approach for pyloromyotomy has no significant influence on operating time or length of recovery
• Laparoscopic pyloromyotomy results in significantly less post-operative discomfort
• Fewer episodes of emesis and doses of Tylenol
• Laparoscopic pyloromyotomy results in obvious cosmetic benefits
ASA, 2006ASA, 2006
• Fibrinolysis had been shown to be better than chest tube drainage alone
• Primary thoracoscopic debridement had been shown to better than tube drainage alone in several retrospective studies
• At the initiation of this study, there were no comparative data between primary thoracoscopic debridement and fibrinolysis as initial treatment for empyema in children
Treatment Of Empyema
• Using our own institution’s retrospective data on length of hospitalization after intervention between thoracoscopic debridement and fibrinolysis with an alpha 0.05 and power of 0.8
• Sample size of 36 with 18 in each arm
Sample Size
Empyema Study ProtocolFibrinolysis
• 12 Fr tube placed by IR or surgery in procedure room
• 4mg tPA in 40ml NS given into tube on insertion and each day for 3 doses
ThoracoscopyThoracoscopy• Thoracoscopic debridement with chest tube left
behind on – 20 cm H20 suction
J Pediatr Surg 44:106-111, 2008J Pediatr Surg 44:106-111, 2008
• Length of hospitalization after intervention (tPA or thoracoscopic debridement) until discharge criteria met (chest tube removed, afebrile & oral analgesics)
Empyema Study ProtocolEmpyema Study Protocol
Primary Outcome Measure
J Pediatr Surg 44:106-111, 2008J Pediatr Surg 44:106-111, 2008
• Days of Tmax > 38CDays of tube drainage
• Doses of analgesia
• Days of oxygen requirement
• Hospital charges after intervention
• Procedure charges
Empyema Study ProtocolEmpyema Study ProtocolSecondary Outcome Measure
J Pediatr Surg 44:106-111, 2008J Pediatr Surg 44:106-111, 2008
Patient Variables at ConsultationPatient Variables at Consultation
WBC WBC 20.820.8 19.719.7 0.71 0.71
Weight (kg) Weight (kg) 24.624.6 20.720.7 0.52 0.52
Age (Years) Age (Years) 4.8 4.8 5.2 5.2 0.770.77
Days of SymptomsDays of Symptoms 9.0 9.0 10.610.6 0.320.32
VATSVATS tPAtPA P ValueP Value
O2 support (L/min)O2 support (L/min) 0.81 0.81 0.79 0.79 0.96 0.96
Study Results
ER/PCP visits ER/PCP visits 2.9 2.9 2.7 2.7 0.69 0.69
J Pediatr Surg 44:106-111, 2008J Pediatr Surg 44:106-111, 2008
OutcomesOutcomes
16.6% failure rate for fibrinolysis16.6% failure rate for fibrinolysis
VATSVATS tPAtPA P ValueP Value
PO Fever (Days)PO Fever (Days) 3.1 3.1 3.8 3.8 0.46 0.46
O2 tx (Days) O2 tx (Days) 2.25 2.25 2.33 2.33 0.89 0.89
LOS (Days)LOS (Days) 6.89 6.89 6.83 6.83 0.960.96
Patient ChargesPatient Charges $11,660 $11,660 $7,575$7,575 0.010.01
Analgesic dosesAnalgesic doses 22.322.3 21.421.4 0.90 0.90
Study Results
J Pediatr Surg 44:106-111, 2008J Pediatr Surg 44:106-111, 2008
London Prospective TrialVATS v Fibrinolysis w/UrokinaseVATS v Fibrinolysis w/Urokinase
• No difference in LOS (6 v 6 days)
• No difference in 6 month CXR
• VATS more expensive ($11.3K v $9.1K)
• 16 % failure rate for fibrinolysis
Am J Respir Crit Care Med 174:221-227, 2006Am J Respir Crit Care Med 174:221-227, 2006
• There appears to be no therapeutic or recovery advantages to thoracoscopic debridement compared to fibrinolysis as the primary treatment for empyema
• Thoracoscopy results in significantly higher patient charges
CONCLUSIONS
J Pediatr Surg 44:106-111, 2008J Pediatr Surg 44:106-111, 2008
PRCTs Now Enrolling• Burn study – SSD vs collagenase (100/150) – Stopped early at interim
analysis
• One stage vs 2 stage laparoscopic orchiopexy for intra-abdominal testis (28/30)
• Standardized feeding protocol vs ad lib feedings following laparoscopic pyloromyotomy (100/150)
• Esophago-crural sutures vs no sutures at laparoscopic fundoplication (both groups receive minimal esophageal dissection), APSA 2010
• Irrigation/suction vs suction alone in patients with perforated appendicitis American Surgical 2012?
• Use of US vs landmark guided CVL placement (with Stanford) (40/80)
• IR drainage of appendiceal abscess with vs w/o instillation of fibrinolytic agent (tPA) (39/62)
• SSULS cholecystectomy vs 4 port lap cholecystectomy, APSA 2012?
• SSULS splenectomy vs 4 port lap splenectomy (6/30)
SummarySummary
• Evidence based decision making will continue to have a stronger presence in medical training
• Patient management will continue to become more influenced by evidence over opinion
• Hurdles to performing prospective trials in surgery and medicine are mostly surmountable
Top Related