Transcript Dr. Alan Forster Feb 19th SSI call

Interactive voice response
systems: A potential method to
track surgical site infections
Alan J. Forster MD FRCPC MSc
Scientific Director, Clinical Quality and Performance
Management, The Ottawa Hospital
Career Scientist, Ontario Ministry of Health and Long
Term Care
Acknowledgments
•
•
•
•
•
TOH
OHRI
CPSI
Capital Health
Vocantas Inc
•
•
•
•
Carl van Walraven
Natalie Oake
Alison Jennings
Nadea Saikaley
Surgical site infections
• Infection at site of surgery within 30 days of
surgical date (1 year if foreign body)
• Common type of hospital acquired infection
– 3rd most common HAI
– Risk 2%-10%
– 30% of SSI involve deep structures
• Costly
• Timing: the majority of SSIs occur after
discharge
Importance of day surgery in
Canada growing
2.5
+16%
+30%
2
1.5
95-96
05-06
-20%
1
0.5
0
Outpatient
No. of cases in millions;
Inpatient
Total
Source: CIHI 2007
Interactive voice response systems
(IVRS)
• Information technology – connects people to
database using a telephone interface
• Consists of specific hardware and software
• Call scripts are programmed
• These are activated by various triggers
• Responses to prompts result in data fields being
populated
• Subsequent prompts (or other actions) occur as a
result of new data
• Commonly used in business
IVRS enables feed-back loops
• Reminders
– Clinic appointments
– Testing
• Monitoring
– Condition
– Compliance
Systematic review of published studies
-Methodology
• MEDLINE search
– Time period: 1950 to January 18, 2008
– Terms: » interactive + voice + response
» calling + system
» (telephony or telecommunications) + voice
» automated + telephone
• Inclusion criteria
– RCTs and CCTs that examined the effect of an IVRS intervention on
clinical and/or process outcomes
• Two reviewers independently abstracted study data
• Outcomes were grouped into 1 of the following categories
–
–
–
–
Clinical endpoints (e.g. hospitalization, death)
Surrogate outcomes (e.g. HbA1c, total cholesterol)
Patient process adherence (e.g. immunization)
Patient-reported quality-of-life (e.g. Addiction Severity Index)
Literature validating patient
response
• > 60 studies have compared the validity and reliability of data
collected using an IVRS with standard data collection
methods (e.g. in-person interviews, paper-and-pencil
questionnaires)
• Examples: » Reported alcohol consumption / substance abuse
» Depression screening tools
• Overall, IVRS yields valid and reliable data
– Cost and time-effective
– Participation rates are generally high
– Participants may be more willing to disclose sensitive
information to an IVRS
Systematic review of published studies
-Results (preliminary findings)
• 38 included studies (29 RCTs and 9 CCTs)
– Median sample size: 230 (Interquartile range 122-648)
– IVRS aimed at changing behavior: N=23
• E.g. immunization, physical activity
– IVRS aimed at managing chronic disease: N=15
• E.g. diabetes, heart failure
• Clinical endpoints – 3 studies
– IVRS intervention associated with improved outcomes in 2 studies
• Surrogate outcomes – 7 studies
– Overall, the IVRS interventions were associated with non-significant
improvements in all outcomes
• Patient process adherence – 28 studies
– Overall, the IVRS interventions were associated with improvements in
outcomes
(median effect 7.85% absolute improvement, 95% CI 2.8-19.5)
• Patient-reported quality-of-life – 8 studies
– Overall, the IVRS interventions were not associated with improvements
Systematic review of published studies
-Conclusions
• IVRSs are increasingly being used in healthcare
settings
• Some IVRSs are part of an overall strategy while
others are stand alone interventions
• Patient process adherence and quality-of-life
outcomes are the most frequently reported
• Few studies examine clinical endpoints or surrogate
outcomes
• IVRSs are a potential solution for improving the quality
of ambulatory care
Objectives:
•To determine acceptability and feasibility of IVRS based
follow up system
•To determine number of patients in whom intervention
based on IVRS changes treatment
•To determine adverse event frequency and type following
discharge
(Am J Manag Care. 2008;14(7):429-436)
Intervention
Hospital
IVRS
WWW
IF YES
Nurse
RN enters Pt data
into IVRS
• Pt ID
• Phone #
• D/C date
Post D/C
Home phone
Cell phone
Other phone (work,
family, friend, etc.)
Questions
1.
2.
3.
4.
5.
Right Pt?
New or worsening symptoms?
(Problems with surgery?)
Problems with medications?
Desire connection to Health Link?
Methods
• Prospective cohort design
• Patients
– Surgical: Consecutive women planned for
gynecological day-surgery
• Exclusion: No phone, dementia, failure to
provide informed consent
STUDY INTERVENTIONS
PATIENT FLOW
Pt admitted to RAH
Gyne Day Surgery Unit
Consent for
study
Info entered into study DB
& Pt assigned study ID
Gyne Surgery
Study ID, phone #, & D/C
date entered into IVRS
D/C from RAH
Post D/C
Day 1
Routine Care
IVRS calls Pt
30 days
Analysis of call
results
Telephone survey
Capital Health DB access
• Encounter Hx
• Lab data
Health Link DB access
• Call records
Analysis of outcomes
Analysis of Pt experience
Study flow
Potentially
eligible patients
N=317
Excluded patients:
N=47 (15%)
Missed opportunity
Refused consent
Patients enrolled:
N=270 (85%)
Entered into IVRS
N=266 (99%)
30 day follow up complete
N= 249 (92%)
Patient characteristics
• Surgery
– Age: 38 years (31-48)
– <10% had a chronic illness
– 4% had no functional limitations in terms of
ADL’s
– Discharged on day of discharge – 244 (98%)
Adverse events
• 33 patients (12% (95% CI: 8%-17%)
• Preventable but almost no ameliorable
• Major types: Therapeutic errors, adverse
drug events, other
IVRS call flow
• 381 automated calls recorded
• 163 patients ‘answered’ calls
• 130 patients (52% of all patients) answered ‘yes’
when asked whether they were the correct
patient
• 129/130 patients answered at least one question
• 104/130 patients answered all questions
• Answered ‘yes’ to at least one question=17
Patient perceptions
n=96
Adverse events
• 40 patients (16% (95% CI: 12%-22%)
• Major types: Wound infection, pain,
bleeding, UTIs, Anaesthesia complications
• 31 patients required corrective actions for
their AE
• Timing of event start: median delay 9 days
(IQR 3.0-15.5)
– 90% of problems started after day 1
Interventions
• No IVRS call resulted in the provision of
HealthLink advice!
• Most AEs happened after call
• Therefore, IVRS did not result in the
identification of any AEs
Summary
• Patients accept IVRS intervention and find
it useful
• System design issues need to be
addressed particularly,
– Timing
– Response
• Efficient method of providing follow up
calls
Proposal: Ambulatory care SSI
surveillance using an IVRS
• Automated calls days 5, 15 and 30 post
discharge
• Simple questionnaire
– Do you have any of the following … at your
surgical site?
– Have you seen an MD following your surgery?
– Would you like to speak with a nurse?
• Notification to nurse working in surgical
unit (PAU, SDCU)
If you do decide to
implement, you need to
consider implementation
barriers.
Consider technical and
organizational factors.
Technical factors
• Patients do not dislike the program
– It is relatively familiar technology and simple
to use
– Often patients think it is a person calling
– Operate on KISS principle
• Software and hardware
– Well established
– System design is relatively straight forward
– System integration is not difficult
Organizational factors
• Financial barriers
– In-house expertise?
• Technical barriers
– IS/IT Departmental protocols
– Ongoing support
• Administrative barriers
– Competing priorities
Summary
• Surgical site infections
• Description of IVRS technologies
• Literature review of IVRS technologies in
healthcare
• Review study using IVRS technologies to
monitor adverse events following day
surgery
• Proposal: can an IVRS be used to monitor
SSIs? - Yes