Return on Investement: Is it possible to calculate Lessons from real life''

1
Return on InvestementIs it possible to
calculate?Lessons from real life..

• Joshua Lee, MD
• Physician Lead, Information Systems
• UCSD Healthcare

2
Overview

• Current methodology in calculating return on
  investement in both order entry and online
  documentation
• Case studies in cost reduction/revenue generation
• Brigham and Womens Hospital
• Dartmouth Hitchcock Medical Center
• UCSD Healthcare

3
Traditional ROI Methods/Asssumptions

• Reducing error through computerized or der enty
  reduces cost
• More appropriate ordering results in reduced
  pharmacy costs
• Drug substitution
• More appopriate dose/frequency
• More online documentation results in reduced
  transcription costs and reduced medical records
  personnel expenditures

4
Error Prevention and Cost

• Most work has been done to demonstrate reduced
  error and hence reduced adverse event from
  implementation of CPOE
• Estimates of increase in LOS by 2.2 day and
  expense by 2,012 for each adverse drug event
• Use of a simulation model preventing errors at
  prescribing, transcription and administration
  showed aggregate reduction in ADEs by 26

Bates, D et al. The Costs of Adverse Drug Events
in Hospitalized Patients JAMA 277 307-17,
1997. Anderson, JG. Evaluating the Potential
Effectiveness of Using Computerized Information
Systems to Prevent Adverse Drug Events. JAMIA,
228-232, 1997
5
Ways in which to project savings

• From "Will Electronic Order Entry Reduce Health
  Care Costs?" Birkmeyer, C. and Lee, J. Effective
  Clinical Practice, March/April 2002. 567-74.

6
Nephros Evaluating Efficacy

• Attempt to have order entry take into account
  altered renal function
• Undertaken at Brigham/Womens long track record
  in CPOE
• Outcome measures
• Appropriateness of dose with adjustment in
  either dose amount or frequency
• Length of Stay
• Total Hospital and Pharmacy Costs
• Use of a cross over design so no longer a
  projection but rather prospective design
  intended to see if proper orders resulted in
  improved cost utilization

7
Nephros Process Outcomes

• 97,151 orders written on 7,490 patients with
  known renal failure
• Of those 14,440 orders with at least one
  modification of either dose or frequency

Chertow, GM and Lee, J et al. Guided Medication
Dosing for Inpatients with Renal Insufficiency.
JAMA. 286(22) 2839-44, 2001.
Plt0.001 for both
8
Nephros Clinical OutcomesChertow, GM and Lee, J
et al. Guided Medication Dosing for Inpatients
with Renal Insufficiency..

• Overall length of stay reduced from 4.5 days to
  4.3 days (statistically significant at p.009)
  but clinically relevant?
• No reduction in overall hospital costs or
  pharmacy costs
• Challenge of ascribing reduction in costs as
  even if one drug cost reduces, need to show
  reduction in overall pharmacy expenditure/patient

9
Med Substitution for Cost SavingsMA Fischer, DH
Solomon et al. Arch Intern Med. 2003 163
2585-2589

• Attempt to measure if pre-programmed alert to
  covert patient from IV to bioequivalent oral
  dosage would effectively change practice
• 11.1 decrease (p.002) reduction in defined iv
  daily dose
• 3.7 (p.002) increase in defined po daily dose
• No sig change in overall pharmacy costs

10
If not OE what then?

• Reduction in hard costs through online
  documentation
• Need to demonstrate first true reduction in
  transcription lines and hence costs
• Need to account for additional hardware, training
  costs and loss of productivity with new
  deployment of documentation tool

11
Deployment of mobile wireless laptops at
Dartmouth-Hitchcock

• Small pilot of 5 surgeons, each provided with
  mobile wireless laptop to replace desktop
• VS/Meds placed by ancillary staff
• Analysis only of notes online
• Laptops twice the cost of desktops
• Full FTE to support wireless program
• But, once in place, refusal to return to prior
  workflow!

89
22
12
What is real institution to do?

• Medical groups and hospitals need to make IT
  decisions based on increasing public demand for
  electronic records and evidence of OE
  implementation
• Immature market and lack of prospective data
  makes these decisions fraught with conjecture
• Need a reproducible strategy that takes into
  account unique features of home institution and
  the expected benefits of the new information
  system/electronic record

13
How we did itA UCSD Case Study

• 400 bed tertiary care beds across two hospitals
• 300,000 annual ambulatory visits in 7 sites, gt30
  clinics
• 268 FTE and 500 total physicians
• Recognized need for a comprehensive ambulatory
  electronic health record
• Needed to justify budget for State of CA
• Needed to justify expenditure by med group

14
Step I AssumptionsCosts/Resource Identification

• Compile all human resource elements
• Physicians by raw number and FTE (price dependent
  on billable FTE)
• Clinical support staff
• Administrative support staff
• Compile costs of dictation, new manual chart
  creation, ongoing chart maintenance, ongoing
  chart movement costs
• Determine ongoing hardware, software and
  operations human support needs
• Allows for projected costs by month of
  deployment then with adoption the expected
  savings accrued

15
Step IIHow long does it take?
Adoption curve valid if rate of adoption either
by successive clinic deployment Or by successive
feature deployment notes, orders, charges.
16
Step IIIMatrix of cost balanced with savings

• With each successive month, does adoption of that
  part of the EMR result in a hard savings?
• Example if 10 of 268 FTEs (25) are using the
  online note editor with savings of 3/visit with
  avg 24 visits/day? 7525 (1875/day or
  56K/month!). Then need to merge with outlay
  costs of the HER.
• Reduction in transcription does not always
  translate into staff reduction
• New roles for medical records and billing
  abstraction personnel

17
Step IVAggregate Savings
18
Step VAggregate Costs
19
Step VIDetermine what NOT to measure

• The vast majority of expected cost savings are in
  the realm of process improvement
• Very SOFT to measure, but in the realm of
• Improved test ordering
• Reduced phone time for ancillary staff
• Pay for performance is next on horizon
• UCSD chose to include neither in ROI

20
Step VIIDo final balances

• What is the year in which expenditures are
  superseded by savings?
• What is the net present value of the project?
• The result of (Savings-Costs) adjusted for
  current rate of inflation
• What are the relative contributions of hospitals
  and medical groups?
• Accrual of benefits not always equal
• Recouping transcription expense accrues to
  medical group.
• Increased compliance laboratory ordering
  augements both professional and technical fees
  (pro fee to med group, facility fee to hospital)

21
Aggregate CostsWhat you take to finance
committee.
22
Take Home Messages

• The current data on true savings is soft at best
• Process improvements do not always translate to
  cost savings
• The best calculations come from careful
  compilation of data about your institution
• IT is increasingly the expected Cost of Doing
  Business