In your Discussion post:
- Explain technology forecasting and its importance
- Differentiate between its different levels
- Discuss the different forecasting methods
- Identify the organizations that are devoted to healthcare technology assessment
- Lead a discussion of any challenges associated with technology forecasting.
Make sure to include citations from your research including this week's content and additional resources you access
Cite powerpoint
Chapter 7
The Policy Analysis Process – Evaluation – Technology Assessment
1
Why Technological Assessment Is Important
Heavy investments in R&D, especial genomic and molecular biology
Application of IT to delivery of care and epidemiology
Increasing market competition globally
Need for better dissemination
Need for wider use of better practices
2
Importance of Technological Forecasting -2
Long research and regulatory lead times
Slowness of adoption due to:
Bureaucracy
Decentralized decision making
Diffusion of power
Need to control costs
Importance of health care to overall economic well being
3
Terminology
Washington State health care technology is “medical and surgical devices and procedures, medical equipment and diagnostic tests”
Let’s include pharmaceuticals here.
Information technology is increasingly a focus of attention.
Little said about organizational innovation.
4
Gilfillan’s Levels
A specific invention
Alternative inventions for the same purpose
Technical accomplishment
Social and economic effects possible from a sets of accomplishments
Social and economic effects likely to flow
Secondary or indirect effects
5
6
Not Just What, But When
In health care one also has to take into account:
Regulation
Rates of dissemination and adoption
which are also affected by coding & payment
_ Further issues of ease of use and compliance
Disruptive Technology
Christensen, Bohmer et al. HBR article
“Health care may be the most entrenched, change averse industry in the U.S. The innovations that will turn it around are ready, in cases – but they cannot find backers.”
IT is one of their examples. In 40 years hospital IT failed to live up to its promise. External backers abound – Leapfrog Group, Newt Gingrich, President Bush, Hillary Clinton, IOM.
8
Disruptive Technology Approach
In large, mature market, go after a segment that can be met with simple, inexpensive technology.
Meet that need, even if the segment is small, but then adapt product to expand into larger segments served by the over-designed product.
Stuff designed at academic medical centers is usually overkill for other places.
9
Supports for Disruptive Innovation in Health Care
Less investment in high-end technology
Simplification of complex problems
Creation of new organizations to do disrupting
Overcoming regulatory barriers
Leadership favorable to change
Per Christensen et al. (2000)
10
Forecasting Methods
Gathering expert opinions
Time series analysis
Surveying and sampling
Correlational and causal modeling
Simulation and system modeling
Integrated approaches
11
Gathering Expert Opinion
Necessary in totally new situations
Can be influenced by acknowledged leaders and dominant personalities
Combat by using variants of Delphi or nominal group techniques
Use the web or LAN to maintain anonymity
Experts tend to favor current paradigm and may be last to see shifts
12
Time Series Analysis
Fit a series of lines to existing data
Identify which line is best fit for the data
Use that line to extrapolate to estimate of the variable being forecast
13
Sampling and Surveying
Sampling behavioral or physical responses, e.g., a clinical trial
Survey people about how they would respond to a real or hypothetical situation, e.g., as in market research
Both are costly and applicable to currently experienced situations.
14
Correlation and Causal Modeling
Works best with a causal model in mind, but can also generate hypotheses with data mining
Example given is the learning curve which is shown to work in cardiac surgery Models can be simple or multidimensional
Greater power as data bases are built and can be accessed
15
Simulation and System Modeling
Suitable for complex situation with interacting (concatenating) variables
Spreadsheet models – look at Figure 10-4
Probabilistic models – Monte Carlo simulations
Feedback models (analog or digital) including systems dynamics
Markov models – usely transitional models with populations moving from state to state
Scenarios – again interacting events
16
Integrated Approaches
Used for many large studies in which we have epidemiological data, multiple diagnoses and treatments, sensitivity analyses, and complex systems
Example is RAND study of future health needs and status of the elderly and of interoperability of health information systems.
17
Required Skills Sets
Identifying population appropriately
Forecasting impact of proposed treatment on THAT population
Projecting uptake of the technology by providers, payers and patient groups.
Estimating the comparative impact of the uptake including secondary effects
Planning release of info to support desirable implementation decisions.
18
Segmentation
Health care is segmented into over 500 products and thousands could be justified. You have to know epidemiology and use that to segment available data to model disease processes, treatments and outcomes.
Example given of alternative t-cell count segmentations and their impacts on cost analysis
19
Organizations Doing This
U.S.
Federal –AHRQ, NIH
State – CA, WA, OR, MN
Private – universities, think tanks, consultants
Other countries
UK – NICE
Canada
The Netherlands, Denmark
20