Core Measure #11 is "Implement Clinical Decision Support".
An example might be an electronic "pop up" that notifies the physical therapist to perform a Falls Risk Screening for Medicare-eligible patients.
Many providers will scramble to implement the "core" measures without fully considering the costs and risks involved.
Here are the risks to using computerized clinical decision support systems (CDSs) in physical therapy:
- Computerized systems can disconnect us from the source of our data. Consider a physical therapist who enters a numeric self-report score from the OPTIMAL scale without first quizzing the patient on high-scoring items, like Completely Unable to Kneel 5/5.
- Computerized systems can cause us to limit our search for data. This fallacy is not limited to CDS systems but is typical of the confirmation bias commonly seen in healthcare settings. Consider the physician who orders an MRI to visualize a lumbar disc in the case of chronic lower back pain but fails to ask about depression.
- Computerized systems can disable the intuition of skilled, experienced decision makers who become accustomed to letting the system make all the decisions.
- Computerized systems can slow the rate of intuitive learning for new users of the systems (e.g.: new PT graduates) so that it takes longer to build intuitive skills.
- Computerized systems can teach dysfunctional skills that actively interfere with learning how to make better decisions. For example, a busy therapist who is paid on a productivity model tries to quickly enter data into her handheld device without conscious reflection or consideration of the data and the resulting CDS recommendations. Do the recommendations make sense?
- Computerized systems use an algorithmic, computer logic that humans may be unfamiliar with. Algorithms, like Treatment Based Classification, may hide the story about how the computer “thinks” about our data. Computer logic is not obvious or intuitive. Computer logic may not match our traditional mechanistic models of human function and pathology.
- Computerized systems have special needs. According to Gary Klein, author of Sources of Power…
“…machines need precise, accurate control and information and we tailor our jobs to meet the needs of machines…”If we are spending our time with the patient hunched over the keyboard then we can be sure we are serving the needs of the machine but not the patient.
- The computerized clinical decision support logical rules become “institutionalized,” rigid behaviors that may eventually have no further bearing on the outcome.
An example of an institutionalized rule is the physician certification of the plan of care.
At one time in the United States, physicians legitimately directed the patients’ physical therapy plan of care. Now, with the exception of post-surgical patients, physicians cannot claim a body of professional knowledge that improves upon physical therapists’ decisions.
- Pop-up fatigue occurs when the CDS delivers excessive “pop-up” windows to the user’s screen during access to the patient record or to the user’s cell phone via text messaging or e-mail.
One study found that 49-96% of alerts were overridden or ignored due to pop-up fatigue. Setting alert triggers to “high severity/critical alerts” can reduce the number of alerts (increased specificity). An example might be an alert that is triggered if the patient’s follow-up functional scores worsen by an amount greater than the MCID/MDC for that test.
- Multi-tasking degrades human performance especially for the group known as heavy media multi-taskers. These people may attempt to carry on a cell phone conversation, text message and send an e-mail simultaneously. While they may feel like they perform each task at the same time, high-resolution, functional Magnetic Resonance Imaging scans reveal that their brain actually switches back-and-forth among different activities. This ability is, appropriately, known as task switching.
Two-hundred and sixty two students were segregated by their media use into heavy media multi-taskers (HMM) and light media multi-taskers (LMM). The students were tested for their ability to filter out irrelevant stimuli and for their ability to task switch. In filtering ability, the HMMs were 77ms slower than the LMMs in filtering out irrelevant stimuli.
In task switching ability the HMMs were 426ms slower than LMMs in switching tasks.
“These results suggest that heavy media multi-taskers are distracted by the multiple streams of media they are consuming or, alternatively, that those who infrequently multi-task are more effective at volitionally allocating their attention in the face of distractions......(HMMs) may be sacrificing performance on the primary task to let in other sources of information.”Since the primary task is the care of the patient in front of the physical therapist an awareness of the danger posed by heavy media multitasking with a CDS system seems imperative.