Class 808H Surgery Workshop

1
Class 808HSurgery Workshop

• NSQIP

2
Advances and PerspectivesThe NSQIPA National
Metric for Quality in Surgical Care
Kamal MF Itani, M.D.,Chief of Surgery,Boston VA
Health Care System
3
Public Law 99-166
December 1986

• The VA should report its surgical outcomes in
  comparison to the national average
• The VA should report its surgical outcomes with
  risk-adjustment, accounting for the severity of
  patient illness

4
OCTOBER 1, 1991 DECEMBER 31, 1993
The VA National Surgical Risk Study for the
development and validation of risk-adjustment of
outcomes models for use in the comparative
assessment of the quality of surgical care
5
QUALITY ?
6
OUTCOME
7
THE NSQIP DATABASE

• 53 Preoperative variables
• 10 demographic
• 30 clinical
• 12 laboratory
• 15 intraoperative variables
• 15 clinical variables
• 34 Postoperative variables
• 10 laboratory variables
• 30-day postoperative mortality
• 21 categories of 30-day postoperative morbidity
• Length of hospital stay
• Long term survival (BIRLS)

PATIENTS UNDERGOING MAJOR SURGERY(First 40
operations in an 8-day cycle)
8
THE NATIONAL SURGICAL QUALITY IMPROVEMENT PROGRAM
(NSQIP)
9
Assurance of Data Quality/Reliability

• Dedicated SCNR
• Centralized nurse training
• Standardized protocol
• Definitions Committee
• Mandatory web-based competency tests
• Hotline to address nurse questions from the field
• IRR assessed periodically by traveling SCNRs
• Site visits on request

10
Top 10 (of 36) Preoperative Predictors of 30-Day
Mortality All Operations
A Beta Coefficient is calculated for each
predictor variable in the model
11
(No Transcript)
12
(No Transcript)
13
(No Transcript)
14
Multivariate preoperative predictors of 30-Day
postoperative mortality and morbidity are
identified annually by logistic regression
analysis
Model

15
Changes in Hospital Ranks After Risk Adjustment
for 30-day Mortality
16
NSQIP Annual Report Mortality O/E Ratios for All
Operations
Statistically significant high outlier
(inferior performance)
Statistically significant low outlier
(superior performance)
17
Risk-adjusted outcomes
NSQIP
Surgical Service
18
Trends in Unadjusted 30-day Mortality Rate for
Major Non-Cardiac Surgery (All Operations)
National Surgical Quality Improvement Program
(FY1992-FY2007)
19
Trends in Unadjusted 30-day Morbidity Rate for
Major Non-Cardiac Surgery (All Operations)
National Surgical Quality Improvement Program
(FY1992-FY2007)
20
Feedback to Providers and Managers

• Annual comparative reports
• Audits
• Instruments and personnel for site visits and
  program reviews
• Observational studies

21
Low outlier facility AMortality
22
High outlier facility AMorbidity
23
Audit- Facility A

• gt9 times the national average for reported
  disseminated cancer.
• 3 times the average for Dyspnea with exertion
• Twice the average for Functional Health Status
  prior to surgery.
• ASA 3 was reported at a much higher rate than the
  rest of the nation.
• Twice the national average for reported wound
  infections in clean cases for FY06

24
Facility B High SSI rate 6.7 vs. 2.3

• It was noted by the reviewers that in addition
  to misclassification of wounds, certain
  procedures such as cystoscopy or incision and
  drainage of an abscess were erroneously ascribed
  a surgical site infection
• Infection control works in isolation and does not
  track surgical site infections across all
  surgical specialties. It was also noted by the
  reviewers that there was an excellent mechanism
  to detect patients colonized with MRSA but that
  no uniform protocol or clinical pathway existed
  for handling these patients in view of the high
  MRSA surgical site infection
• Prophylactic antibiotics are not dosed based on
  patient weight. Blanket dosage is given to all
  patients.
• Prophylactic antibiotics are not uniformly
  re-dosed in long cases
• Clipping of hair is done in the OR.
• Compliance with hand washing, especially alcohol
  based scrub, in the OR is not monitored.
• Cultures from surgical site infections and
  organ/space abscesses are not uniformly
  performed

25
Solution- Facility level

• Continuous education
• Focused reviews
• Heightened awareness to status compared to
• National average (ASA, SSI, MI)
• Yearly trends
• Proper follow-up of patients

26
Feedback and site visits

• Outlier Hospitals

27
High Outlier Level of concern triggering a paper
audit or site review in the NSQIP
The NSQIP Board might perform site visits for any
level based on available data or other Criteria
available to the board at that time.
28
Site visits

• Structures and coordination of care

29
Organizational StructureStructural
parameters assessed by the NSQIP in high and low
outlier hospitals

• Technology and equipment
• Technical competence
• Interface with other services
• Relationship with affiliated institutions
• Monitoring of quality of care
• Coordination of work
• Leadership
• Overall quality of care

Daly et al, JACS 1997 185341
30
The Role of Coordination
Feedback
31
Best practices for low outliers

• Utilization of a variety of coordination
  mechanisms
• Effective use of peer interaction
• Regular meetings of surgical service leaders to
  address administrative issues
• Feedback based on objective data
• Development and utilization of clinical pathways
• Strong surgical service leadership
• Provision of adequate resident supervision

32
Observational studiesThe Case Of Colorectal
Surgery

• Learning from the data

33
Goals

• To understand the structures and processes of
  care that might contribute to a higher mortality
  in patients undergoing elective colon and rectal
  surgery in the Veterans Affairs (VA) Hospital
  System
• To examine various domains of care each patient
  that died after surgery in facilities detected as
  high outliers on unadjusted and adjusted
  mortality
• To assess the impact of volume of colon and
  rectal surgeries on mortality in the VA Hospital
  system

34
(No Transcript)
35
Methods

• All NSQIP colon and rectal surgeries between
  2003-2005
• Open procedures
• partial colectomy (44140-44141, 44143-44146,
  44160)
• abdomino-perineal resection (44147),
• total colectomy (44150-44156)
• mobilization of splenic flexure when performed in
  conjunction with partial colectomy (44139).
• Laparoscopic procedures
• Partial colectomy (44204-44208)
• Total colectomy (44210-44212)

36
Methods

• The crude 30 day mortality for the whole VA
  System and each hospital separately were
  calculated
• VA facilities with a 30-day mortality rate double
  the national average were selected for review.
• Models were used to develop the predictive models
  for colorectal surgery deaths using an available
  59 variables for each patient. Variables that had
  bivariate associations with 30-day mortality
  (plt0.20) were submitted for consideration in a
  stepwise logistic regression.
• Ratios of observed deaths (O) to Expected deaths
  (E) were then computed for each hospital.
• A 90 confidence interval was computed for each
  O/E .
• Confidence intervals which showed that a facility
  is a significantly high outlier (plt0.10) was
  selected for review

37
Methods

• NSQIP review questionaire, operative report,
  pathology report, autopsy, medical record.
• Colorectal malignancy
• metastatic disease,
• emergency surgery,
• delay in diagnosis, delay to surgery,
• suboptimal surgical procedure,
• system issues and practioner care issues.
• Peer review by facility
• Reconsideration by NSQIP board
• A Spearman rank correlation test between facility
  volume of colorectal surgery and colorectal
  unadjusted mortality over the years 2003-2005 was
  performed

38
Results
39
Results
40
Episode of care- examples

• Delay in surgery
• Surgery delayed for 24 hours or greater on a non
  decompressed large bowel obstruction in an acute
  care bed.
• Deteriorating patient with clostridium difficile
  colitis taken for surgery after developing
  hemodynamic instability
• Appropriate surgery
• Radical resection performed in patients with
  diffuse liver metastasis and multiple co
  morbidities
• Four hours attempt at laparoscopic resection
  followed by conversion in a 94 year old patient
  with multiple co morbidities

41
Episode of care- examples

• Delay in Diagnosis
• Colonoscopy delayed for months up to a year after
  detection of heme positive stools
• Delay in diagnosis of a large bowel obstruction
  or sigmoid volvulus
• System issues
• No mechanisms in place for follow-up on
  colorectal screening and referral
• No mechanisms in place for follow-up on abnormal
  laboratory or radiologic tests
• Practitioner care issues
• No attempt to improve hematocrit preoperatively
  in an anemic patient with coronary artery disease
  undergoing major resection
• No deep venous thrombosis prophylaxis in a high
  risk patient who died from pulmonary embolus

42
(No Transcript)
43
Conclusion

• The majority of colon and rectal deaths occurred
  in older patients with malignant disease.
• In colon and rectal surgery the episode of care
  starting with timely diagnosis, timely referral
  to surgery, timely surgery, choice of the
  operation, and post operative care are all
  important in defining the final outcome of the
  patient after surgery.
• Volume of colon and rectal surgery cases did not
  correlate with the outcome

44
Total Costs for Major General-Vascular Surgery
Procedures According to Specific Complications
(VA BHCS)
Compared to the group with no complications
45
Cost for 1513 Patients in Whom Process
Improvement Effected a Change in the O/E Ratio
From 1.20 to 0.80
46
Ann Surg 2005242326-341
47
Effect of Specific Complication Groups on Median
Survival of the Total Study Population
48
New Features of NSQIP

• NSQIP Website Dashboards
• The NSQIP website https//vhadennsqipweb.v19.med.v
  a.gov/NSQIP.aspx
• Dashboard for registered users
• Status of NSQIP case assessment workload
• Mortality morbidity O/E ratios for the most
  recent reporting period
• Unadjusted mortality morbidity rates
• Links to historical current quarterly and
  annual reports

49
VISN Dashboard Report
https//vhadennsqipweb.v19.med.va.gov/NSQIP/Defaul
t.aspx
50
New Features of NSQIP

• 6-Month Rolling Reports
• Starting with Q1 of FY08
• O/E ratios are calculated for the current quarter
  for the last 6 months
• All subspecialties have O/E ratios presented
• 6-month O/E ratios are more likely to detect
  significant issues than one quarter of data alone

51
NSQIP Dashboard Reports
https//vhadennsqipweb.v19.med.va.gov/NSQIP/COSumm
ary.aspx
52
Coming Soon to NSQIP Website

• Index Operations
• Commonly performed surgeries in each subspecialty
  
• Colectomy surgery piloted, with O/E ratio and
  unadjusted mortality outliers detected
• Other planned surgeries
• Lung Resection
• Non-Ruptured AAA
• Knee Replacement
• Others TBD
• Will be added to an NSQIP dashboard

53
NSQIP OR Self Assessment Safety Reports
(LIST TRUNCATED)
130 VAMCs 5 observational reviews/month 99/130
(76) compliant
https//vhadennsqipweb.v19.med.va.gov/NSQIP/SelfAs
sessmentVISN.aspx
54
How to Use NSQIP Information

• View NSQIP dashboards when new reports are
  released
• Last week of March, June, September for quarterly
  reports
• February for annual reports
• Review your O/E ratios unadjusted outcomes
• indicates significantly high O/E
• indicates significantly low O/E
• Look at each specialty is there a potential
  problem in a particular specialty?

55
The Institute of Medicine has described the
quality of care in the VA as the best in the
nation, citing the VA NSQIP as one of 3 major
factors that have contributed to improved quality
of care.