"Physical therapy is not a subspecialty of the medical profession and physical therapists are not medical doctors; we are a separate profession that provides a unique service that physicians are unable and untrained to provide."

Letter to the AMA from the APTA, Dec 2009

Thursday, August 26, 2010

What Kind of Decision Maker are You?

Jill was a bright new PT graduate with a quick smile and a winning personality. She had worked her tail off in physical therapy school and she had a head full of knowledge she was ready to use.


Jill was especially keen on treatment based classification and new clinical decision rules. Her training surpassed the education of many of her older peers, though, and she was sensitive to their professional pride when making her treatment recommendations. Nevertheless, Jill was "on fire" to use her new skills and she did so with gusto - impressing her patients and the rest of the staff with her authority and her good results.

Bill was Jill's boss. Bill was a seasoned expert in many therapy settings - currently he worked as Director of Rehab in a hospital outpatient department and treated patients about half of each day.

Bill had also learned about the new clinical decision rules through some directed self-study and had used the rules on some of his patients.

After 20 years of treating patients, however, Bill felt that he could do just as good as the rules in predicting treatments - he had even subjected his judgement to his own little test.

He evaluated some patients with his judgement and then measured them using the rule - he found that his judgement matched the rule almost all of the time.

Bill had seen many new graduates and he recognized Jill's enthusiasm but he also noticed that she seemed to have something different from the other new graduates he had mentored - more than just enthusiasm and intensity - Jill also had a systematic approach to measuring her patients and making decisions.

Their differences came to an impasse when Jill requested that Bill create new, computerized templates for the hospital electronic medical records (EMR) program. Jill needed specific outputs, such as expected frequencies, duration and outcomes based on her patients' individual data.

Jill had been entering her data into the EMR but the data just sat there - nothing was printed on the Plan of Care that went to the physician for signature. Jill wanted the EMR to automatically interpret her data based on existing decision rules and make recommendations. Jill had to manually enter her recommendations using free text typing which took valuable time away from patients.

Bill knew that Jill's request would be problematic:
  1. software coding for the EMR would cost money,
  2. no therapist consensus existed on the need or the efficacy of TBC,
  3. the literature on TBC was incomplete
  4. and many, experienced staff would resist changing their documentation habits based on the recommendation of a new graduate.
What Should Bill do?

Jill uses quantitative models to make her decisions. Quantitative decision making is on the rise in healthcare - although providers, especially physicians and physical therapists - still have a ways to go in improving our decision making fidelity.

Bill, however, uses qualitative decision models that are the hallmark of experienced professionals in many fields.  There is substantial evidence that physicians use qualitative over quantitative decision models. According to the University of Texas Medical Informatics Department physicians...
  1. have difficulty with quantitative reasoning
  2. have difficulty diagnosis using Bayesian analysis (making diagnoses based on prevalence, test results and posterior probabilities)
  3. have difficulty interpreting effectiveness of treatments
  4. have difficulty estimating probabilities (and, as a result, infrequently use probabilities in practice)
As a result of these deficits in quantitative reasoning physicians may...
  1. order excessive, expensive and invasive diagnostic tests
  2. incorrectly interpret the test results
  3. inconsistently interpret the post-test probabilities of disease
  4. make inconsistent treatment decisions
  5. and over-treat conditions with infrequent poor outcomes
Decision researchers usually contrast quantitative vs. qualitative decision making although decision researcher Gary Klein, in his book Sources of Power, uses the term "naturalistic" instead of qualitiative.

Naturalistic decision making (also called "pattern recognition") has also been criticized over the last 25 years in decision research as computers and "computer-like" decision algorithms have become more popular.

Klein argues that both models are helpful and uses the metaphor of peripheral and foveal vision to illustrate that naturalistic decision making is a "wide angle" approach that captures all relevant (and some irrelevant) data while quantitative decision making (TBC/CDR) is a "narrow" approach that captures only relevant data and rules out all other options.


Klein presents the case that naturalistic decision makers in fields as diverse as...
  1. US Navy missle defense and flight commanders
  2. Firefighters
  3. Chess grandmasters
  4. Smokejumpers
  5. Nuclear power plant risk managers
  6. Software designers
  7. Corporate CEO's
...do not use quantitative models for over 90% of their decision making tasks. Instead, these experienced experts rely on naturalistic decision making to manage their day-to-day tasks.

Which is Better?

The current healthcare crisis may imply that "something" needs to be done and perhaps improving our decision making models will improve...
  1. costs
  2. outcomes
  3. medical errors
  4. provider liability to audits
  5. provider liability to medical malpractice
  6. efficient allocation of societal resources
There is good evidence that costs, outcomes and medical errors can be improved using CDR. Current data suggest that the medical error rate across various settings and geographic regions are similar:

RegionError Rate
Great Britain3.7%
Colorado2.9%
Utah3.7%

Error rates in industries that have implemented computerized clinical decision support, however, are markedly different:

IndustyError Rate
Airlinesless than 0.01%
Bankingless than 0.01%

What Did Bill Do?

Bill could see the writing on the wall - he knew the day of pure naturalistic decision makers in healthcare - the old guard who relied on "gut instinct" and experience to provide care - was coming to an end. Cost pressures and the enthusiasm of people like Jill would usher in a new dawn that used computers and algorithms for the simple decisions. He hoped he would still have a role to play.

When Bill watched Jill at work he felt better - if the future depended on people like her then he knew that physical therapy was in good hands.

2 comments:

  1. I know! I want a system that does just as you are suggesting! The current electronic systems for physical therapists have the capability to not only capture the data but also help analyze the data. When I talk to vendors, the typical response I get, "no one else is asking for that."

    Why not some system that can give ideas/suggestions at the end just before saving? Why not a system that can share how similar patients did AND what happened during those episodes of care? (Not only for me the clinician, but for me to share the info with the current patient! Throw in the actual cost of the service too to help with financial decisions for the patient.) Also, there are known medications with side effects that will affect outcome or maybe THE reason for the problem. Why not some capability for a system to screen the meds and combination of meds and a pop up if an issue is apparent?

    Also, less is more. We really can make good clinical choices with less information. Not every single patient we treat is complicated. Some are and an algorithm will be of no help... but the more data we have the better we can be at making decisions versus using our memory.

    If you haven't seen this yet, I was intrigued... Phil Gabel has meshed the Orebro Musculoskeletal Questionnaire with functional tools. I think you'd be interested in checking out his system and talking to him. He uses an algorithm to predict return to function using the above tools.

    You can do a scholar google for "CP Gabel" and you'll see the research he has published in the area of outcomes. Those tools are the ones used in the software program he designed.

    The program is web-based, so you just log in to use it. You can find it here: http://www.adviserehab.com/

    There is also a quick explanatory YouTube video on the software here: http://www.youtube.com/watch?v=vRexn1dJ2G0

    Quite interesting and it actually excites me to see tools like the above used in algorithms.

    ~Selena

    ReplyDelete
  2. Selena,

    I'm also exciting about Dr. Gabel's system. I DO think clinicians are asking for these features/benefits - but maybe the vendors are not understanding.

    For example, one of the most pressing needs in outpatient physical therapy is the need to predict the duration and cost of the physical therapy Plan of Care. Dr. Gabel's system does this - so do pen-and-paper based TBC templates for a variety of common, high-volume conditions.

    One of the problems with Clinical Decision Support systems is that, from 1994 to 2005, their success rate at altering clinician behavior has DECLINED. I believe this is because they are becoming more complicated.

    As a result, computerized CDS systems may overpromise and underdeliver.

    Dr. Gabel's system looks very simple and straightforward. I think it has potential to deliver as promised.

    Tim

    ReplyDelete

Free Tutorial

Get free stuff at BulletproofPT.com

Tim Richardson, PT owns a private practice at Medical Arts Rehabilitation, Inc in Palmetto, Florida. The clinic website is at MedicalArtsRehab.com.

Bulletproof Expert Systems: Clinical Decision Support for Physical Therapists in the Outpatient Setting is a manager's workbook with stories, checklists, charts, graphs, tables, and templates describing how you can use paper-based or computerized tools to improve your clinic's Medicare compliance, process adherence and patient outcomes.

Tim has implemented a computerized Clinical Decision Support (CDS) system in his clinic since 2006 that serves as a Reminder, Alerting, Prompting and Predicting CDS using evidence-based tests and measures.

Tim can be reached at
TimRichPT@BulletproofPT.com .

"Make Decisions like Doctors"


Copyright 2007-2010 by Tim Richardson, PT.
No reproduction without authorization.

Share PTD with your Peers!

American Physical Therapy Association

American Physical Therapy Association
Consistent with the American Physical Therapy Association Vision Statement for Physical Therapy 2020, the American Physical Therapy Association supports exclusive physical therapist ownership and operation of physical therapy services.