Showing posts with label EHR. Show all posts
Showing posts with label EHR. Show all posts

Saturday, October 5, 2019

Physicians Preservation Act?






I got this idea today while reading the usual Twitter complaints about the electronic health record. A post by physician I knew was particularly poignant. She pointed out that she was getting burnout from the excessive time it takes to do EHR documentation compounded by the fact that nobody ever reads it. This is a complaint I have had for a long time.  I was lucky enough to be on the ground as the EHRs rolled out. There was quite a prelude to the rollout with about a solid 10-year buildup of documentation and billing requirements. Those requirements originated with the federal government specifically HCFA – the precursor to CMS. All of the initial EHRs were designed around these documentation and coding templates. It was strictly a business focus sold as something necessary for medical practice.

I can recall the first people on the medical staff who were designated to sell the system. They came to see me and I pointed out that I have never been a touch typist and the fastest I can type is 12 words a minute with two mistakes. When they realized I wasn’t kidding they tried to soften the blow by saying that we would be slowly transitioned to creating the entire document. During that transition time we would still be able to dictate admission notes and discharge summaries. When I complained that this would still be quite a burden on physicians producing all these documents I was told by an internist (who I had a very high opinion of) “You need to thank our CEO for getting us this state-of-the-art system.” That was one of the more depressing remarks that I’ve heard in my career.

I did try to make the most of it. I got an early version of Dragon and started dictating all my notes and into Word and pasting them into the EHR.  It was not pretty. There were many mistakes and if I missed some of those mistakes it could prove to be an embarrassing document. The nursing staff I worked with helped to edit those documents and point out the mistakes but some mistakes invariably went through. I learned that the nursing staff in my immediate proximity were the only people who ever read those notes. I was generating multiple 500 to 1000 word documents a day and suddenly realized that I had to complete that work between 10 PM and midnight every day. Within a few years the new car smell was off the EHR and things were getting ugly.  I started to see 18 to 20 page progress notes based on import and cut-and-paste features. My speculation is at one point the vendor was desperate to prove they could introduce some physician friendly features. The ability to start a new daily progress note based on yesterday’s note soon became history. Administrators decided that the new note looked too much like the old note even though they were based on same template.

EHR politics is always interesting to observe. There are a cadre of administrators and “super users” who are tasked with selling the product to the frontline physicians. There are also various helpdesks that are run by the vendor. Staff at those helpdesks are supposed to be available for troubleshooting and problem solving. The troubleshooting and problem solving eventually fades away. EHRs are typically implemented in modules. I walked into work one morning and realized that the module that allowed electronic prescribing was completely changed. The change was not announced and since it was an enterprise wide implementation there were hundreds of physicians trying to figure it out for themselves. It added hours to everyone’s day.

With the shift of billing, coding, and documentation to physicians many other jobs were lost due to the EHR. For 15 years I would go to the basement of the hospital every Sunday and make sure all of my records were dictated and signed. I ran into the same staff there every weekend who greeted me and assisted me with completing those records. Suddenly they were gone because now I was doing all of their work in the EHR. When I first started working at my job, I would dictate daily progress notes and they would be pasted into the chart by the secretarial staff the next day. Billing and coding specialists would come to the unit, read those notes, and attach a billing fee. I had no idea about the billing system and didn’t really care. With the EHR all of those staff were replaced. I was not only doing their jobs but now I was legally responsible for any billing errors and the suggested penalties were high. All of this additional work and responsibility was directly transferred to physicians through the EHR.

The only real bright spot from the EHR was the ability to see imaging studies, electrocardiograms, and laboratory results as soon as they were available. It took years to get that implemented to the point it worked effectively.

Are there workarounds to successfully use the EHR without burnout, depression, and excessive work? I think that there are. The last few years I have been seen by ophthalmologists who were retinal specialists and an otolaryngologist or ENT physician. In both cases these positions were using a scribe or a third person in the room who documented the history, exam, findings, and treatment plan as indicated by the physician. In the case of the retinal specialist he was working with an ophthalmology fellow and made corrections to that examination by directions to the scribe. The same thing happened with the ENT physician but in that case the scribe was also an RN who could provide more details about the suggested treatment plan. In both cases the physician walked out of the room at the end of the encounter with no further documentation burden. That led one of my colleagues to point out that the only reasonable workaround for the EHR problem is to use two people - the physician and a scribe or staff person who could also function as a scribe.

That led to my idea about the Physicians Preservation Act at the top of this post. It addresses all the flaws in the system that were brought about by heavy lobbying and Congressional advocacy for a burdensome inefficient electronic documentation system. As I pointed out in a 2015 post, the system has never lived up to claims of efficiency or savings even when physicians started to do the work of four or five people. This entire administrative structure is there to produce excessive documentation that nobody reads. There is also a massive environmental cost since the system must operate through thousands of networked personal computers that in many cases are operating 24/7 along with the associated data storage facilities. 

My suggested solution is a compromise between the likely inertia of the current EHR system and the politics that keep it in place and the massive burden it places on physicians and their families. There are just too many special interests in Congress keeping this system afloat. The question is how long can the country afford to lose doctors because of it.

My guess is not too much longer.


George Dawson, MD, DFAPA


Sunday, December 9, 2018

What Isn't Available In Multimillion Dollar EHRs? Decision Support from 1994


Physician Decision Support Software from the 20th Century




I used to teach a class in medical informatics. My emphasis was not mistaking a physical illness for a psychiatric one and also not missing any medical comorbidity in identified psychiatric patients.  The class was all about decision-making, heuristics, and recognition of biases that cause errors in medical decisions. Bayesian analysis and inductive reasoning was a big part of course. About that time, software packages were also available to assist in diagnostic decisions. Some of them had detailed weighting estimates to show the relative importance of diagnostic features.  It was possible to enter a set of diagnostic features and get a listing of probable diagnoses for further exploration. I printed out some examples for class discussions and we also reviewed research papers and look at the issue of pattern recognition by different medical specialists.

The available software packages of the day were reviewed in the New England Journal of Medicine (1).  In that review, 10 experts came up with 15 cases as written summaries and then those cases were cross checked for validity and pared down to 105 cases.  The four software programs (QMR, Iliad, Dxplain, and Meditel) were compared in their abilities to suggest the correct diagnosis. Two of programs used Bayesian algorithms and two used non-Bayesian algorithms. The authors point out that probability estimates varied based on literature clinical data used to establish probabilities. In the test, the developers of each program were used to enter the diagnostic language and the compared outcomes were the list of diagnoses produced by each program. The diagnoses were rank ordered according likelihood.

The metrics used to compare the programs was correct diagnosis, comprehensiveness (in terms of the differential diagnosis list generated), rank, and relevance.  Only 0.73-0.91 of the programs had all of the cited diagnoses in the knowledge base. Of the programs 0.52 - 0.71 made the correct diagnosis across all 105 cases and 0.71-0.89 made the correct diagnosis across 63 cases.  The 63 case list was used because those diagnoses were listed in all 4 knowledge bases.  The authors concluded the lists generated had low sensitivity and specificity but that unique diagnoses were suggested that the experts agreed may be important. They concluded that the performance of these programs in clinical settings being used by physicians was a necessary next step. They speculated that physicians may use these programs beyond generating diagnoses but also looking at specific findings and how that might affect the differential diagnosis.

A study (2) came out five years later that was a direct head-to-head comparison of two different physicians using QMR software to assess 154 internal medicine admissions where there was no known diagnosis.  In this study physician A obtained the correct diagnosis in 62 (40%) cases and physician B was correct in 56 (36%) of the cases. That difference was not statistically significant. Only 137 cases had the diagnosis listed in the QMR knowledge base. Correcting for that difference, correct diagnoses increased to 45% for physician A and 41% for physician B. The authors concluded that a correct diagnosis Listed in the top five diagnoses 36 to 40% of the time was not accurate enough for a clinical setting, but they suggested that expanding the knowledge base would probably improve that rate.

Since then the preferred description of this software has become differential diagnosis generators (DDX) (3.4). A paper from 2012, looked at a total of 23 of these programs but eventually included only 4 for in their analysis. The programs were tested on ten consecutive diagnosis-focused cases chosen from from 2010 editions of the Case Records of the New England Journal of Medicine (NEJM) and the Medical Knowledge Self Assessment Program (MKSAP), version 14, of the American College of Physicians. A 0-5 scoring system was developed that encompassed the range of 1= diagnosis suggested on the first screen or first 20 suggestions to 5= no suggestions close to the target diagnosis. The scoring range was 0-50. Two of the programs exactly matched the diagnosis 9 and 10 times respectively. These same two programs DxPlain and Isabel had identical mean scores of 3.45 and were described as performing well. There was a question of integration with EHRs but the authors thought that these programs would be useful for education and decision support. They mention a program in development that automatically incorporates available EHR data and generates a list of diagnoses even without clinician input.

The most recent paper (4) looked at a a systemic review and meta-analysis of differential diagnosis (DDX) generators. In the introductory section of this paper the authors quote a 15% figure for the rates of diagnostic errors in most areas of medicine. A larger problem is that 30-50% of patients seeking primary care or specialty consultation do not get an explanation for their presenting symptoms. They looked at the ability to generate correct lists of diagnosis, whether the programs were as good as clinicians, whether the programs could improved the clinicians list of differential diagnoses, and the practical aspects of using DDX generators in clinical practice. The inclusion criteria resulted in 36 articles comparing 11 DDX programs (see Table 2.)  The original paper contains a Forest Plot of the results of the DDX generators showing variable (but in some cases high) accuracy but also a high degree of heterogeneity across studies.  The authors conclude that there is insufficient evidence to recommend DDX generators based on the variable quality and results noted in this study.  But I wonder if that is really true.  Some of the DDX generators did much better than others and one of them (Isabel) refers to this study in their own advertising literature.

My main point in this post is to illustrate that these DDX generators have been around for nearly 30 years and the majority of very expensive electronic health record (EHR) installations have none.  The ones that do are often in systems where they are actively being studied by physicians in that group or one has been added and the integration in the system is questionable.  In other words, do all of the clinical features import into the DDX generator so that the responsible clinician can look at the list without making that decision.  At least one paper in this literature suggests that eliminates the bias of deciding on whether to not to make the decision to use diagnostic assistance.  In discussion of physician workflow, it would seem that would be an ideal situation unless the software stopped the workflow like practically all drug interaction checking software.

The drug interaction software may be a good place to start. Some of these program and much more intelligent than others. In the one I am currently using trazodone x any serotonergic medication is a hard stop and I have to produce a flurry of mouse clicks to move on.  More intelligent programs do not stop the workflow for this interaction of SSRI x bupropion interactions.  There is also the question of where artificial intelligence (AI) fits in.  There is a steady stream of headlines about how AI can make medical diagnoses better than physicians and yet there is no AI implementation in EHRs designed to assist physicians.  What would AI have to say about the above drug interactions? Would it still stop my work process and cause me to check a number of exception boxes? Would it be able to produce an aggregate score of all such prescriptions in an EHR and provide a probability statement for a specific clinical population?  The quality of clinical decisions could only improve with that information. 

And there is the issue of what psychiatrists would use a DDX generator for?  The current crop has a definite internal medicine bias.  Psychiatrists and neurologists need an entirely different diagnostic landscape mapped out.  The intersection of psychiatric syndromes, toxidromes and primary neurological disorders needs to be added and expanded upon. As an experiment, I am currently experimenting with the Isabel package and need to figure out the best way to use it.  My experimental paradigm is a patient recently started on lithium who develops an elevated creatinine, but was also started on cephalexin a few days after the lithium was started.  Entering all of those features seems to produce a random list of diagnoses and the question of whether an increasing creatinine is due to lithium or cephalexin.  It appears that the way the diagnostic features are entered may affect the outcome.

Decision support is supposed to be a feature of the modern electronic health record (EHR). The reality is the only decision support is a drug interaction feature that varies greatly in quality from system to system. Both the drug interaction software and DDX generators are very inexpensive options for clinicians.  EHRs don't seem to get 1990s software right.  And it does lead to the question: "Why are EHRs so expensive and why do they lack appropriate technical support for physicians?" 

Probably because they were really not built for physicians.



George Dawson, MD, DFAPA


References:

1: Berner ES, Webster GD, Shugerman AA, Jackson JR, Algina J, Baker AL, Ball EV, Cobbs CG, Dennis VW, Frenkel EP, et al. Performance of four computer-based diagnostic systems. N Engl J Med. 1994 Jun 23;330(25):1792-6. PubMed PMID:8190157.

2: Lemaire JB, Schaefer JP, Martin LA, Faris P, Ainslie MD, Hull RD. Effectiveness of the Quick Medical Reference as a diagnostic tool. CMAJ. 1999 Sep 21;161(6):725-8. PubMed PMID: 10513280.

3: Bond WF, Schwartz LM, Weaver KR, Levick D, Giuliano M, Graber ML. Differential diagnosis generators: an evaluation of currently available computer programs. J Gen Intern Med. 2012 Feb;27(2):213-9. doi: 10.1007/s11606-011-1804-8. Review. PubMed PMID: 21789717.

4: Riches N, Panagioti M, Alam R, Cheraghi-Sohi S, Campbell S, Esmail A, Bower P.The Effectiveness of Electronic Differential Diagnoses (DDX) Generators: A Systematic Review and Meta-Analysis. PLoS One. 2016 Mar 8;11(3):e0148991. doi: 10.1371/journal.pone.0148991. eCollection 2016. Review. PubMed PMID: 26954234.




Wednesday, August 1, 2018

The Problem With Checklists.....





I have critiqued the checklist approach to psychiatry in many posts on this blog.  Several like-mind psychiatrists have also added many comments in this area. I had recent experience with surgical checklists that leave a lot to be desired.  So much so that if I was not an MD - I might not be sitting here and typing this post right now.  For now, I will just post the bare bones sequence of events for illustrative purposes.  On April 14, I had an operative procedure that required antibiotic prophylaxis that consisted of a single intravenous dose of antibiotics given right before surgery. On July 31, I had a second operative procedure to address complications of the first procedure. Both procedures were done under general anesthesia (fentanyl and midazolam or Versed).  A laryngeal mask airway (LMA) was used instead of intubation.  The general sequence of events went like this.

1.  Preop physical exam - good for 1 month prior to surgery.  The exam is done by a primary care MD.  The surgery will not occur without it.  The goal is to identify and complicating or potentially contraindicating conditions.  Specific instructions are given to the surgeon and anesthesiologist based on this assessment.  Specific instructions are given to the patient about if they need to change up their medications at all prior to the surgery.  For example, it is common advice to hold aspirin and other anti-inflammatory medications (NSAIDs), and certain vitamin supplements for 1-2 weeks prior to the surgery.

2.  Hospital intake - over the hour or two before surgery there are intense meetings with a number of disciplines:

Pre-Op RN:  Reviews the medication list and confirms that all of the recommended medications and pre-op instructions were followed.  Assesses functional capacity as well as presence of eyeglasses, hearing aids, implants, pacemakers, CPAP mcahines and artificial joints.

Pharmacist:  Reviews the detailed list of medications and looks for potential drug-drug interactions as well as drug-anesthesia interactions.

Anesthesiologist:  Reviews the detailed list of medications and rationale.  Takes a detailed cardiovascular history. Examines heart and lungs.  Asks detailed family history and personal history for anesthesia interactions particularly malignant hyperthermia. In both cases this hospital trained nurse anesthetists who asked the same questions and administered the pre-op midazolam before leaving the pre-op area.

 OR Nurse:  Also interviews patient about concerns over the surgery and assures that all intravenous lines and devices that will be in the operating room (OR) are present and working.

That is the overall sequence of events.  Each of these team members has specific jobs and checklists that were entered into an EHR.  The primary care physician handed me a copy of my pre-op exam to take with me in case the faxed version was lost.  It was printed out from a well known enterprise wide EHR.

I have a condition called lone atrial fibrillation that is commonly seen in middle-aged (and now old) men who exercise too much.  It was originally thought to be associated with high levels of dynamic exercise like cycling and running, but epidemiological studies suggest it may also be associated with jobs that require a lot of heavy lifting - like furniture and piano moving.  I have also talked to power lifters in the gym who developed it when they continued the lifting into their 50s and 60s.  I take flecainide and it keeps me in a steady sinus rhythm and that has worked well for the past 8 years.  The problem with flecainide is that it is a fairly toxic medication if you have the wrong biological substrate or if you mix it with the wrong medications. A trial of flecainide in ventricular tachycardia was halted because of increased fatalities in the treatment group compared to placebo.  The last electrophysiologist I talked with suggested that I needed to get an exercise stress test done every year to make sure that the QRS interval was not widening due to the drug.  For the purpose of these surgeries my primary concern was not getting a medication that would interact with flecainide and result in a fatal arrhythmia.  I knew that this surgical specialty used fluoroquinolones preoperatively and if you search that interaction in any database this is a typical result.           

"Moderate risk - can cause QTc prolongation and should be avoided when possible. Increased risk for torsades de pointes and other significant ventricular arrhythmias.  Other factors (old age, female sex, bradycardia, hypokalemia, hypomagnesemia, and higher concentrations of the interacting drugs can increase risk for potential life-threatening arrhythmias."

I naturally wanted to avoid the fatal arrhythmia.

At every step in the above chain, I explained this drug interaction and advised the team members that I can safely take cephalosporins.  And here is what happened.

In both cases I had the same primary care MD doing the pre-op physical exam.  He was very focused on the pre-op checklist and in fact the rooming medical assistance reviewed the med list, vital signs, and review of systems that was entered into the EHR checklist before I saw the physician.  When he was done he asked me if I had any concern and I told him "Any antibiotic or anesthesia agent that interacts with flecainide - I do not want to take. I know that I can take fentanyl and Versed for general anesthesia so those are the preferred agents if they can use them for this surgery."  The first time he pulled up the interaction in the EHR, agreed and said - "I will flag this in my assessment so they see it."  The second time he said the same thing but reviewing the H & P he handed me it was not present.  It is possible it was transmitted on another form.

And so it went with every members of the preop team.  They all seemed surprised every time I brought it up.  They thought I was talking about an allergy as opposed to a drug-drug interaction. One of the pharmacists looked it up on her Smartphone app and confirmed.  There was a lot of confusion about the preop antibiotic right up until the time of administration.  Was there another drug that could be used? Would the doctor change the standard orders to administer another drug?  For the past surgery - I had to tell them to look up the April record and confirm that I was given 2 grams of IV cefazolin and not levofloxacin.

When it was finally clarified, it took two nurses to figure out how the levofloxacin could be discontinued from the standard order in the EHR so that the cefazolin could be given.  I was finally given the cefazolin, operated on and so far (barring another complication) things are going well.

The lessons:

1.  Almost everything you hear about the EHR and checklists increasing safety is a myth -  

In this case all of the professionals were using state of the art (and extremely expensive) EHRs containing checklists and forms that were dutifully completely and the ultimate check here was the patient who happened to be a physician who compulsively studies drug interactions and cardiac complications.  That is not a level of safety that I want to see for any of my family members or patients who are undergoing surgery.

2.  Patients or competent family members are the best safeguards for safety at this point -     

I have worked with very bright and insightful nurses who told me that they have a rule that they accompany hospitalized colleagues and check everything that is going to happen to them as well as what medications are administered to them.  On the other hand I have asked patients what medications they take and been told: "You tell me doc.  I just put them in a pill container.  I don't know the names, doses or what they do."

There is a lot of talk about patient empowerment, but it has to be built on a solid foundation of patient literacy.  I certainly realize that a lot of people do not want to know, but I have also talked with many people having less than a high school education who could tell me every drug they took during a complicated course of cancer treatment that included a bone marrow transplant.  Reading and understanding the pharmacy printout given with a medication is a basic prerequisite for the literacy I am talking about. 

3.  This is a systems problem and not a personnel problem -  

Let's face it - all of the personnel in the system are highly competent licensed professionals. They are all focused on their tasks and they do a good job of it.  The problem is that all of these very competent individual assessments are not synthesized into a useful safety plan.  Experts have been writing about the importance of checklists in industry (like the airline and automotive industries) for decades but medical information is individualized complex, and not redundant.  Any adverse outcome of the sequence of events that I described is likely to be something like this:

"Well Dr./Nurse X - did you fail to read all of the narratives in the chart and the patient stating that he had concerns about drug interactions with flecainide?"

Any response to the effect that the EHR is difficult to read and in some cases incoherent and should have flagged this concern automatically is likely to be met with:

"Well that's our EHR and we have to live with it. Our focus groups with the nurses and the manufacturer have been working on fixes for the past 5 years."

Translation:  somebody has to take the blame and it won't be the EHR.



4.  Why doesn't the EHR/checklist approach work? 

It failed miserably - not just once but twice in my case and I was advocating at every level to flag flecainide and not give me any interacting drug.  Having worked with EHRs now for nearly 20 years I can speculate on a few things.  First, there is very little intelligence built into EHRs.  In this case the EHR will do a comprehensive drug interaction search on the current list.  But there is probably not an automatic search on the standard preop antibiotic.  If there is - physicians are numbed to dismissing so many false positive drug interactions that could have happened as well. Second, any discussion of the patients concern or doctors advice is buried in documentation that is prioritized for billing, rarely read, and not translated into any rational action. An intelligent EHR would convert the concern about flecainide interactions into what is called a hard stop. That means the potentially offending drug could not be ordered until some further action was taken - like a discussion between the physician, pharmacist, and patient.  In this case, my discussion with 10 people was not beneficial and the only reason I did not get levofloxacin was that I was in a hospital bed about 6 feet away from where the nurse was working and I was a physician who has worked for years to prevent these kinds of problems.

It is hard to believe that such extremely expensive and heavily lobbied systems can't provide a basic level of safety.  I was not surprised to read that having the same primary care physician for years is probably a better assurance of longevity.

For the non-medical person reading this - know your medications, what they do, and what the potential safety concerns are when you are in a situation where those medications are being changed. Ask your pharmacist and physician to do a drug interaction search to make sure these transitions can be safely made. Refuse any medication unless a sound rationale can be provided to you about why you are taking the drug and how safe it is to take with your current prescriptions.


George Dawson, MD, DFAPA

     
Graphics Credit - the graphic at the top is from Shutterstock per their standard licensing agreement.








Tuesday, October 31, 2017

Updated Review of Systems (ROS) for Medical Psychiatrists







The review of systems for psychiatry has changed significantly over the years.  Those changes were due to billing and coding decisions rather than clinical utility.  If you have been practicing long enough you have witnessed the transformation from a document very similar to what primary care physicians use, to one that is more focused on the sleep and appetite disturbances associated with psychiatric disorders. That results in a distinctly different ROS in psychiatry than the rest of medicine.  That puts medically based psychiatrists like myself at a disadvantage because the electronic health record (EHR) templates may not include the physical symptoms that I am most interested in and that requires more documentation. 

These changes are not unique to psychiatry.  Patients find themselves filling out checklists in many clinics that are essentially a surrogate ROS.  Something that your physician used to ask you in person and ask you to elaborate on is now a checklist.  In the modern EHR, the ROS is often just a series of checkboxes.  No elaboration required thank you.  The form that you fill out in the waiting area is incorporated into the physician's note often without reviewing it with the patient.  In some systems employing scribes or persons to do the documentation the scribe will type or dictate this form into the record. There is one additional point where the physician might read any ROS incorporated this way and that is during the read of the final note for signing.  That review is usually cursory because of time constraints - I doubt it is read with any regularity.  There is not enough time to read documentation 2 or 3 times as it is complied, transcribed, and entered into the EHR.

Another EHR strategy that is used from time to time is a statement: "A complete 10-point review of systems was done and it was negative."  Use of that statement depends on the billing, coding, and compliance staff and whether they think it meets the subjective standards of the day to demonstrate to somebody that the work was done. In my experience, unless you are interviewing a very healthy 20 or 30 year old it is unlikely that the ROS is completely negative.

The expanded ROS is more specific to medicine and it assumes that the physician is asking clarificatory questions.  I have found over the years that a very basic structured exercise like the ROS produces very different results depending on asking all of the questions, asking clarificatory questions, and pursuing obvious leads to other sets of questions depending on the patient response. Treating the ROS like it is a static series of yes or no questions is likely to produce the minimum amount of diagnostic information.

As an example consider the following example:

The ROS is being conducted on a 75 year old man.  He is being seen for insomnia.  In taking the medical history he says he was diagnosed with congestive heart failure 2 years earlier. He has impaired physical performance due to CHF and can only walk 100 feet and slowly climb a flight of stairs with great effort. On the ROS he endorses needing to prop himself up to breathe and occasionally wakes up suddenly at night due to shortness of breath.  Those symptoms and additional physical exam findings suggest that CHF is the problem rather than insomnia and the treatment needs to change accordingly. 

A more common example:

The ROS is being conducted on a 50 year old man.  He denies any cardiac or pulmonary symptoms and is only taking an 81 mg aspirin in addition to two different antidepressants. He has a 30 year history of smoking a pack a day of cigarettes.  The interviewer asks: "Have you ever had stress test?"  The patient states he does not know what that is. "You walk on a treadmill and they keep increasing the grade until you have to stop".  The patient replies that he took the test and the Cardiologist came in and sprayed something into his mouth. "Do you think that was nitroglycerin spray?" The patient states that it was and he had an immediate angiography and stent placement. 

Both examples illustrate that the ROS is dynamic and not static.  Filling it out in the waiting room may seem to be efficient, but the amount of information obtained in that setting is likely to be low relative to real medical problems that exist.  The probability of increased information from a more dynamic ROS increases with the age of the patient due to accumulated medical problems with age.

In addition to the list of symptoms in the ROS, additional heuristics at the level of pathological mechanisms can be considered to hone in on a specific syndrome.  The following table illustrates two of them.  For example, the General category in the ROS generally implies some kind of infectious, metabolic or endocrine condition - but it is not enough to make an actual diagnosis without further delineation. ROS categories are supposed to roughly correlate with body systems rather than pathological mechanisms, but many of the symptoms do not have a definable body system.


VINDICATE


VITAMIN D


V – Vascular

I – Inflammatory

N – Neoplastic

D – Degenerative / Deficiency

I – Idiopathic, Intoxication

C – Congenital

A – Autoimmune / Allergic

T – Traumatic

E –  Endocrine



V – Vascular

I – Inflammatory

T – Trauma

A –  Autoimmune

M –  Metabolic

I – Iatrogenic

N –  Neoplasm

D -  Degenerative


If I think the patient has a flu-like illness I ask about specific symptoms of flu-like illness. In addition to fatigue, weight change, fever, chills I might ask about - malaise, cough, rhinorrhea, nasal, congestion headache, sore throat, myalgias, chills, and sneezing.  Positives on several of these symptoms greatly increases the likelihood of a diagnosis of a flu-like illness.  Asking those questions occurs when an infectious etiology is suspected.

The typical review of systems that I used for years is printed below with red highlights for additional points that I ask if there are any markers in the initial history that suggests that they might be positive.  For example, if I am seeing a 50 year old with a long history of stimulant use, on three different antihypertensives and an anti-arrhythmic medication I will generally ask all of the cardiopulmonary symptoms and the additional questions about cardiac testing imaging and diagnoses.  For example: "You mentioned that you have never had a heart attack or a stroke, but has any doctor every told  you that you had cardiomyopathy or a thickened wall of the heart? Do you remember where all of that testing occurred?"




Review of Systems

General:  fatigue, weight change, fever, chills, night sweats

Endocrine: hot or cold intolerance, thyroid problems, hx of neck irradiation

HEENT: decreased visual acuity, hearing loss, tinnitus, vertigo, epistaxis, hoarseness or voice change, sinus/nasal infection or discharge, ear pain, history of ear infections, decreased auditory acuity

Pulmonary: dyspnea, cough, sputum production, chest pain or tightness, hemoptysis, asthma, bronchitis, emphysema, hx pneumonia, hx TB, hx positive/negative PPD, smoking hx
polysomnography, CPAP, APAP, BiPAP, nightmares, night terrors, parasomnia

Cardiovascular: chest pain, palpitations, tachycardia, syncope, edema, orthopnea, paroxysmal nocturnal dyspnea, claudication, phlebitis, hypertension, hx rheumatic heart disease, family hx heart disease
stress test, echocardiogram, angiography, stent placement, congestive heart failure, cardiac ablation, cardiac event monitoring, tilt table testing

Gastrointestinal:  nausea, vomiting, hematemesis, melena, dysphagia, indigestion, heartburn, abdominal pain, abdominal swelling, jaundice, hx hepatitis, hematochezia, diarrhea, constipation, hernia, hemorrhoids, peptic ulcer disease, gallbladder disease, pancreatitis, GI surgery
esophagogastroduodenoscopy, colonscopy, hepatic ultrasound, pancreatitis

Genitourinary: urinary frequency, urgency, dysuria, nocturia, hematuria, hx kidney stones, flank pain, hx STD, genital lesions, testicular mass or pain, sexual dysfunction
Hx acute renal failure

Gynecological: menarche, menopause, last menstrual period, description of menstrual periods, pelvic pain, vaginal discharge or bleeding, sexual dysfunction, breast mass, breast discharge, last breast exam, last mammogram
pregnancy history, hx pre-eclampsia or eclampsia

Skin: mole, other lesion, pruritus, rash, bruises, contusions, lacerations, burns, hx skin cancer

Hematopoietic: excessive bleeding, hx anemia, family history of disorder, lymphadenopathy

Neurological: headaches, migraines, ataxia, incoordination, vertigo, gait problems, falls, loss of consciousness, seizures, head injury, skull fracture, focal weakness, focal sensory change, hx stroke, micropsia, macropsia, metamorphopsia, chronic pain
Brain imaging, EEGs, coma, encephalitis, meningitis, chronic fatigue syndrome, movement disorders

Musculoskeletal:  joint pain, joint stiffness, joint swelling, muscle cramps, muscle pain, muscle wasting, hx fractures
Gout, Lyme Disease, fibromyalgia, rheumatic diseases, treatment by rheumatologist

Allergic/Immunologic: hay fever, rhinitis, seasonal symptoms
Allergy testing, specific allergens, immunotherapy


These are techniques that I have found useful over the years.  In psychiatry, the ROS is useful because I frequently have gotten past the medical history section and inquired about all major surgical and medical diagnoses from the past and the result is surprisingly thin.  More specific prompting about the diagnoses and which physicians the patient has seen in the past can produce much more information in an interview setting.  For psychiatric purposes, the ROS is also included in follow up visits and it seems necessary.  I find it useful for documenting intercurrent illnesses and medication side effects.  

Each class of psychiatric medications has their own relevant ROS that can be recalled with practice.  I might try to type those out at some point in time - but not tonight.  My main point here is that the ROS does have a function above and beyond the psychiatric history for psychiatrists.  People tend to view it as a difference necessary for one billing code or another.

I see it as an opportunity to figure out what is really going on medically with my patient and possibly diagnose another illness. It is also necessary to know that the patient does not have an underlying medical condition or treatment for that condition that contraindicates or necessitates closer monitoring of the proposed psychiatric treatment.


George Dawson, MD, DFAPA


ROS Files:  You can download the ROS files used for this post at the following links as Word documents.  Any suggestions for further modification appreciated:

ROS modified 

ROS standard



Additional Fact: 

A poster on Twitter [Alasdair Forrest @alasdairforrest] let me know that the ROS in the UK is called "systemic enquiry".






Monday, September 18, 2017

Medication Reconciliation


The Medication Reconciliation Process
Medication reconciliation has become a term that is much en vogue after the Joint Commission and electronic health record (EHR) manufacturers got a hold of it.  Medication reconciliation (MR) basically means that anytime a patient changes health care setting there needs to be a procedure in place to assure that their medications are not changed as a result of that transition - no medication errors can be made.  The best example is a patient goes from an outpatient setting into a hospital for surgery, has the procedure done and is discharged.  This care transitions would require a medication reconciliation at the time of admission and another one at the time of discharge.  Prior to the EHR, a physician would just transcribe the admission orders rewriting or modifying the patient's outpatient medication as necessary.  With the EHR, the MR can be rapid if the patient and the medication are already in the system, or it can be a very slow process if they have to all be entered from scratch.

In my 22 years of inpatient experience, I have a lot of experience with this process.  I would be the recipient of shopping bags full of medications and learn that some patients were taking greater than 20 medications at a time.  Most of my time on inpatient units, I was in charge of reconciling all of the patient medications - medical and psychiatric on the patients I admitted and discharged.  In extreme cases this process alone could take an hour on either end.  It got a lot worse over time because more people were inserting themselves in the process.  Pharmaceutical benefit managers learned to demand  an entirely new prior authorization process, even for medications that the patient had been taking for years - at the time of discharge.

The reconciliation process has several modifications based on the care model.  For example, in less acute care settings where physicians are not present in the facility at all hours, on-call staff call in and do remote medication reconciliation.  Before the EHR that would involve a discussion with nursing staff who would review the medication, the patient's status and put in the orders.  The physician would countersign these orders the next day.  In current EHRs, there is a med reconciliation section and there may be an expectation that the reconciliation occurs at the time of admission.  Nursing staff will typically enter the medications from available bottles or pharmacy records.  The physician has to pull up the record remotely, review the patient's status with nursing staff, and sign off on the entered medications.  In some systems, only the basic prescription is ordered and the physician will have to complete numerous fields before the inpatient orders are complete.  Modern EHRs invariably include drug interaction software with very low thresholds and all of those warnings need to be clicked through and dismissed before the patient's usual medications can be resumed.  It is a very slow and inefficient process compared to before the EHR.

One of the rationalizations for the MR in the EHR is patient safety.  Regulatory bodies like the Joint Commission are very big on safety factors and they should be.  The EHR was supposed to greatly reduce errors due to illegible written orders, but in this case the physician was giving verbal orders to nursing staff.  A quick glance at the graphic at the top of the page illustrates some of the thought and decision making that needs to go into this process.  There is really no known way to make it fool proof.  Subjective determinations about medications and medication safety are being made at every step of the way.  Errors in MR still occur largely because of the assessment required by nursing and in this case psychiatry.  The easiest way to conceptualize this is to think about people who take prescriptions and whether they take the medication exactly as it is prescribed on the label.  The commonest problems involve patients taking their medications at the wrong time or all at once.  Often they have been advised by their physician to make a change but the prescription has never changed.  In many cases they have stopped the medication and want to restart it.  The prescribing physician may not know that the patient is using alcohol, other substances, or nutritional supplements with the prescription.  Continuing medications, stopping them or modifying them requires significant clinical judgment and there may be a lot of uncertainty about the history obtained.

Medication reconciliation is a complex and potentially lengthy task.  It works best with experienced nursing staff who can get the best information to the prescribing physicians and then physicians who have good clinical judgment and flexibility to adapt to changing histories.  It is a potential area for artificial intelligence applications.  AI could assist nursing staff without replacing them.  We need an optimal algorithm and a full description of the decision space associated with this process.  Most importantly we need computer applications that support staff rather than getting in their way and requiring staff support of their own.

One of the most interesting aspects of the current conceptualization of medication reconciliation is how it is perceived by administrators and regulators.  The idea that medications can be entered into a piece of software that is essentially a word processor and that makes things right is almost magical thinking to me.  All of the hard work leading to that conclusion (see graphic and beyond) is not only ignored - but nobody seems to get credit for it.  It is the old rationalization: "If it isn't documented it didn't happen."  This is clear proof that most of what happened isn't documented.  If would be impossible to function if it was.  At some level it appears that all of this hard work was produced by the EHR and not physicians and nurses.  Credit to the geniuses who came up with the software and the administrators who decided to put it in.  How did we ever practice medicine without them?

We end up documenting what we are told to document and it is a poor substitute for what actually happened.  Some of the underlying reasons for that documentation are almost always political.    
          


George Dawson, MD, DFAPA

Monday, September 11, 2017

HITECH Editorials in the NEJM...





Since my Labor Day message to colleagues dovetails so well with these editorials, I did not want to miss the opportunity to comment on them.  They appear to be written by people with policy interests in this information technology takeover of clinical medicine.  They are mildly critical but totally miss the mark on what a catastrophe this government roll out has been.  The question any taxpayer should ask is why any other outcome would be expected.  Software and network implementations world wide and at the level of the US government have led to colossal failures.  Multibillion dollar investments  that at some point were abandoned.  The only difference in this case is that the government is not the actual client.  The federal approach to health care - apart from the brief foray of FBI agents raiding physicians offices to see if they made any coding violations is to set up payments for proxies and let them hash things out with providers.  The primitive approach of marginal incentives that are really weighted as penalties is supposed to facilitate the whole mess.  The mess would get implemented either way if you ask me.  There are tens of thousands of executive and mid level health care executives chomping at the bit for a project like this to mismanage.  And they have mismanaged it well.  Government leverage makes it difficult to refuse.

The initial editorial by Washington and  co-authors (1) focuses on the success of getting hospitals and physicians on the electronic health record (EHR).  They present a graphic showing the steep increase in EHR use over the 2004 to 2015 decade.  The acute care hospital curve ends at essentially 98-100% for certified EHRs and office based practices are at 90%.  The article rightly points out that physicians have borne the brunt of the implementation and how physicians are frustrated by the lack of "actionable information generated by these systems".  The article discusses the need for the "seamless flow of electronic information" in a couple of places.  It describes how EHR could be useful in research.  It ends  on a vague note that there is still a lot of work to be done and maybe that will happen some day.

The second piece by Halamka and Tripathi (2) starts out on a more realistic note.  Top down implementation gave physicians inadequate tools and then blamed them for being reluctant.  Technically physicians were not reluctant because they did not have a choice.  In most systems, administrations made all of the purchasing decisions, overhyped the software, and let it be known that contrary opinions were never appreciated.  It was up to physicians to learn how to use the stuff no matter how time consuming it was.  They point out that some measures were enacted on top of the clinical workload that made the situation worse.  They include the longest sentence I have recently seen in a journal article but it does cite a fair number of the problems a lot of the problems:          

"Soon physicians were expected to provide high-quality and empathic care in a 12-minute visit while weaning themselves from paper-based workflows, entering the numerous structured data elements required for meaningful use, rolling out new HIPAA privacy notices, implementing security protections for new electronic data, learning and incorporating new ICD-10 billing codes, and convincing their patients to use patient portals and secure e-mail, all while avoiding safety and malpractice issues." (p 907).

At one point they make the argument that health care organization have moved to "value-based purchasing".  Was that applied to the EHR?  Is there anyone today who would suggest that any EHR that is currently sold in this country is a value based proposition or is there as a result of HITECH legislation?  In their conclusion they suggest that now all of these systems are installed - the government can afford to pull back simplify requirements and let market effects shape some of the metrics like interoperability.  They suggest that returning control to the customers is a path to "recapturing the hearts and minds of our clinicians."

The government heavy aspect of these editorial pieces cannot be denied.  It is more of the same "we are from the government and we are here to help you whether you want us to or not."  Here are a few aspects of this roll out that the HITECH legislation either missed or made a lot worse:

1.  Incredible cost - 

Enterprise wide systems are incredibly expensive both upfront and for the annual licensing and maintenance fees.  That does not include any modification of the system - that will typically cost more.  Once a health system has bought in - it is difficult to shop around and come up with a better deal.  In some areas one company has a monopoly on the enterprise.  In many cases the systems are marketed as being a lot more easy to use than they are.  Support is huge in the implementation phases and drops off in a hurry.  Subsequent modifications - even if they are easy to make cost large sums of money.  In some cases the vendors demonstrate whiz bang technology like seamless integration with voice recognition systems.  The customers often find out that those options don't work well with their systems or are available as a high priced option.

In many organizations the EHR budget (combined with other federal costs cutting measures) is a fixed drain on the budget.  If revenues fall, lay offs can occur just to keep the EHR running.  In private practices, the up front and monthly licensing fees are no less of a burden.  There are some "free" EHRs that are funded by advertising or research but no standard comparison or guidance for any clinic that needs to implement one.  The total budget of these costs would be interesting to see, but I have never been able to find a good reference.  Health systems typically describe their margins in the low single digits.  If that is true and EHR system costing tens of millions up front with tens of millions in maintenance costs is clearly a tremendous drain on the system.

2.  IT implementation is poor -

I don't know what percentage of physicians has seen their EHR rolled out in a way that does not optimize clinical utility.  Working physicians need the most rapid route to incorporate the EHR into their work flow.  That includes software that works, software that is efficient, and ideally software that is smart enough to allow individual physicians to analyze trends in the same patient or groups of patients that will allow better diagnosis and treatment.  The IT implementation is also frequently biased toward administration rather than clinicians.  Many clinicians are surprised to find that someone is counting their mouse clicks as a way to measure productivity and the EHR charts they access are monitored.  This is another significant cost that nobody ever seems to discuss. The most egregious implementation error is when a software change is made on the fly and the physicians are given a heads up with no training.  They are expected to learn the software change with no training.  I have always found the illusion of assistance with the EHR interesting.  For the first few months there are always superusers and the factory reps clamoring to help you out.  They gradually fade into the background and you are left with a very poor piece of software.          

3.  Software quality is poor - 

As far as I can tell current EHR programs are designed to deliver lab and imaging data, generate documentation and reports, and perform a billing and coding function.  They do a fair job with the labs and imaging details. Documentation is very labor intensive and poorly done.  It adds hours per day to the physician's work flow and has necessitated the hiring of scribes and retired physicians just to keep up with the documentation tasks.  It is common that EHRs cannot be accessed by outside physicians and when that happens - the printout sent to those physicians is poorly structured and extremely content poor.

On the authorship side - a basic goal should be to produce a document fairly quickly that appears to have been written by an intelligent being.  As anyone who has read EHR entries or reports that is not typically the case.  There are extremes at either end.  You can find notes that are basically a series of check boxes or you can find 18 page notes where the author imported everything that they could into the note because that is one of the few things (in some EHRs) that you can do quickly.  Neither approach is helpful in terms of continuity of care or developing rational treatment for a patient.  Having used EHRs for the past 15 years - I can attest to clunky editing and incompatibility with voice recognition systems as being major drawbacks.  The text fields of some EHRs only work with their own microscopic and very slow editing tools.  It is impossible to set a cursor anywhere in the field to produce the document.  Using this twenty times a day when you are used to working with functional word processors is maddening.  Some systems of care set a font that looks like it is out of the 1950s and that is how the final document appears.

Every physician was appointed (under penalty of law) to be their own billing and coding specialist.  Sure every hospital and clinic has some billing and coding specialists but today they are there basically to audit the work of physicians.  In the EHR this translates to a tedious search for the diagnoses, listing them in the right priority, and signing off on the diagnostic and billing codes.  This can take up to an additional 20 mouse clicks per encounter.  Even if you can do that in 2 minutes - times 4000 encounters per year - that equals another 133 hours per year. That is work added just to maintain the EHR.  Before the EHR, billing and coding could be completed in about 10% of the time.

All the time physicians are engaged in these inefficient EHR based practices they are hearing how the EHR is such an advance in efficiency and productivity.              

4.  Hardware infrastructure/software is running 24/7 - 

Before the modern EHR, there were a limited number of workstations per hospital and most of them were shut down at night.  Now there are thousands of workstations and storage arrays in large organizations running 24/7.  They can't be shut off because of frequent software updates.  Nursing and medical staff can easily be observed spending most of their workday at computers rather than talking with patients and families.  Before the current EHR, physician would typically look at a computer screen to review the labs and possible the MAR (record of meds given).  Now starting at a computer screen most of the time is the norm.  The EHR dominant approach has increased the electrical bill and reduced time spent with patients at the same time.     

5.  A question of security - 

There have been well publicized leaks of large numbers of patient files and more recent ransomware attacks.  Security in most software systems has historically been an afterthought.  I have not seen any specific problems with EHR software but this tip sheet from CMS points out the potential complexity of the situation.  The security problem is also more urgent for healthcare sites that are under more stringent privacy requirements like 42 CFR Part 2.    

Those are a few of my ideas about the rapid deployment of the EHR.  Unlike the authors I am very skeptical of any drastic improvements on the horizon.  If you can't make an EHR that will produce a coherent report with information content at least equal to an old admission or discharge note that is a major problem.  If you can't produce an EHR that allows for some intelligent analysis of data without going through the entire record and reading every text note that is a major problem.  Sure - access to labs is nice, but we had computer access to labs before the EHR.  Patient access is also nice, but let's be honest - it is limited and doesn't address what patients really want - quality health care.

About the only thing that I agree with the authors on is that the physician needs to be put back into the loop.  But that hides the very basic fact that physicians were intentionally taken out of the loop thirty years ago when politicians decided that they could be replaced by managed care administrators.

When you look at it from that perspective the massive problems with the current EHR - make perfect sense.
  

George Dawson, MD, DFAPA



References:


1:  Washington V, DeSalvo K, Mostashari F, Blumenthal D. The HITECH Era and the Path Forward. N Engl J Med. 2017 Sep 7;377(10):904-906. doi: 10.1056/NEJMp1703370. PubMed PMID: 28877013.

2:   Halamka JD, Tripathi M. The HITECH Era in Retrospect. N Engl J Med. 2017 Sep7;377(10):907-909. doi: 10.1056/NEJMp1709851. PubMed PMID: 28877012.

Friday, September 1, 2017

Happy Labor Day VI



I missed my Labor Day message last year for some reason.  I must have been too burned out.  Burnout has been a big theme in the physician  community in the last several years.  It is almost like it is a new discovery or another new epidemic.  Now we have detailed comparisons of degree of burnout by medical specialty and even some country to country comparisons.  The curious phenomenon about burnout is how physicians are blamed for it.  The typical intervention is to have a "course" on how to "handle" burnout.  You know mindfulness, meditation, yoga, and time management.  There is never any focus on the fact that physicians just work too hard because they have to work free for so many people.  Managed care companies, pharmaceutical benefit managers, government bureaucrats at various levels, and their own employers have come to expect that American physicians have nothing better to do than devote their time and energy to the betterment of those collective businesses.  By that metric physicians are the most exploited employees in the USA.  There is no other group expected to work for so many businesses for free.

Don't get me wrong.  I am not saying that physicians are not paid well.  I am saying that according to the studies I read they are being paid for anywhere between 50-75% of the time they work.  Even the time that they are being paid for is deeply discounted.  What other group of professionals in the USA is expected to work on an arbitrary productivity scale that varies greatly from payer to payer and has a superimposed global budget and federal incentives and penalties superimposed on top of that?  Only physicians work in that environment.  Only physicians are expected to teach future medical professionals for free.  Another one of those cases where the the term "professionalism" can be marched out and used against you.  An example:  "You can just pick up this course for the medical students or residents.  It should not take much time and I know you like to teach.  They want you to teach."  The next several weekends (whether you are on call seeing patients or not) will be devoted to to coming up with PowerPoints.  Lectures and seminars in medicine these days don't happen without the ritualistic exchange of PowerPoints.  In the process PowerPoints get blamed.  I actually like PowerPoints if they are done correctly.  I think they are unfairly blamed when the burden to suddenly produce them is displaced onto the medium rather than the process.

But the focus of my missive today is not burnout or the root cause of excessive uncompensated work.  It is one of the sources of uncompensated work and that the the electronic health record (EHR).  The attitude toward the EHR has shifted in a direction that I have promoted for over a decade.  When the group I was working for was presented with the EHR and trained on it, I knew it was a problem from the start.  We were moving from a hospital wide system that was basically for entry of medication orders by health unit coordinators, vital signs, labs, and nursing notes.  All of the progress notes were dictated or hand entered. At the end of the day I printed out the MAR (record of medications given) for each of my patients and double checked all of the medications they were getting.  The MAR was a single page table showing all of the medications in the left margin and day columns to the right with times of administration.  It was all dot matrix printing - so not as stylish as modern printouts.

Back at the time when politicians were overhyping the EHR and how it would save the health care system hundreds of billions of dollars - I did a little experiment with one of my new hospital admissions.  I decided to read all of the outpatient medical notes to see how many significant medical diagnoses were being carried over in the EHR.  At the time we were online with the new system for about 8 years.  It took me 4 hours to find and read all of the notes from Internal Medicine, Endocrinology, and Cardiology.  There were 236 notes in all. But in the end I noticed that 10 significant diagnoses had been dropped somewhere along the line.  Nobody ever seems to want to acknowledge the complexity in medicine.  As people get older they accumulate an incredible number of medical problems and in some cases the only indication is a very long list of medications that they are taking.  They have been seeing an equally long list of physicians truncating that list of diagnoses because of time constraints.  In the EHR you eventually end up looking at a very short list and need to reproduce a comprehensive evaluation from scratch.  So much for the time savings of the EHR.  Even the politicians are quiet on that one for now.

A new EHR experiment happened to me just recently.  I still treat medically complex patients and often receive them from acute care hospitals where they may have been in intensive care units.  Since many of them were taking various psychiatric medications, I felt obligated to see what the intensivists, cardiologists, pulmonologists, and gastroenterologists all had to say about these medications and whether they complicated critical care or ongoing care of the chronic medical problems.  I want to see the results of ECGs, labs and imaging studies.  I want to know if the patient received any of their usual medications when they were in the ICU or general hospital.  Before there was an EHR all of this information was contained in about 10 very readable pages consisting of the admission note, discharge summary, MAR, and a couple of sheets including the actual ECG tracing and lab reports.

For the event in question I read through the EHR printout.  It was 48 pages long.  It contained limited data.  Blood pressure trends and readings were not printed even though that was one of the critical parameters being followed.  The physician notes were jumbled paragraphs considerably less that traditional reports.  The bulk of those notes consisted of checklists and imported data in different fonts and margins.  The appearance was chaotic.  Who uses 14 point Courier font in documents these days?  I haven't seen that since the days of the telegraph.  In that entire 48 page document there were about 6 lines in a cardiologist's note that made sense so I locked onto those for my report.

I was less optimistic about Phase Two.  I have been working in my current position for about 7 years and during that time I have requested MARs on hospitalized patients about 50 times.  I have received exactly zero.  In this case for some reason it went through.  I received a 60 page fax that was the MAR.  The patient was critically ill and delirious at one point, so there were five different infusions used in intensive care spread out across many of the pages.  The composition of the solutions were listed and the specific rates of infusion.  If I wanted to know the exact amounts that the patient received - it was up to me to figure it out.  Two critical factors from the MAR that were not evident from the EHR printout.  First, the patient was much more critically ill than described in the EHR printout.  Second, none of the maintenance psychiatric medications were given.  Total time to figure all of this out - 60 minutes.  In addition to the read of the EHR printout and interview - total time for the evaluation and report was 2.5 hours.

It is impossible for physicians to do a good job of patient care without all of the material I reviewed in this case.  On the other hand, there are few places in the USA where the physician has 2.5 hours for each new evaluation.  That is how you end up with truncated problem lists, partial medical care, and physicians staying in clinic 3 hours after everyone else has left.  Without the data there is not enough information for the physician to have a decent informed consent based discussion with the patient on the new set of risks associated with a critical illness.

The real culprit here is the fact that physicians have lost control of their profession.  We have had an overhyped, inefficient, ridiculously high-priced piece of software foisted upon us by politicians and the businesses that they support.  It is really no better than personal database software that I was using in the 1990s and that software produced a more readable and coherent report.  The only reason the software works at all is because there are a million physicians out there with work arounds and doing the uncompensated hard work necessary to keep it afloat.

There is no better topic to comment on this Labor Day.  This is my wish to all of my colleagues trying to avoid repetitive stress injuries from the mouse clicks and typing necessary to support EHRs everywhere (I had to switch to my left hand about ten years ago).

Happy Labor Day!


George Dawson, MD, DFAPA          


Attribution:

Picture is Titian's work Sisyphus in the Public Domain from Wikimedia Commons at https://commons.wikimedia.org/wiki/File%3APunishment_sisyph.jpg


         





Saturday, July 30, 2016

The Problem With EHR Software - A Clear Example




The above example is as clear as it gets in terms of illustrating the problem with electronic health record (EHR) software and associated hype and government mandates.  The idea that we need an EHR is a given, and I am not arguing that point.  I am arguing that the current software is inefficient, on par in many ways with software I was using on my PC in the 1990s, high maintenance, and a tremendous burden to any physician who has to use it.  It is also vastly overpriced with no end to that overpricing in sight - largely due to a monopoly of manufacturers and the use of a licensing model for the software.  And like practically every process in medicine these days, the implementation and actual use of EHRs is a highly politicized process that is far removed from the people who have to use it every day.

In the above example, I am tasked with a basic titration of gabapentin according to a recent research protocol (1).  All of the doses used are generic 300 mg capsules of gabapentin for the purpose of simplification.  The dose is titrated over 3 days to 300 mg TID (three times a day) or 600 mg TID.  People reading this may have picked up prescriptions with instructions typed out on the label about how to increase the dose to a therapeutic level.  In settings where a particular medication is used repeatedly and across a large patient population, the rate of titration and capsule side may need to be varied but the concept is the same.  

The question is how do I get this information to the pharmacy so that the medication can be dispensed to the patient in the most effective manner.  In the "old days" of paper records or the early hybrid models where  all of the orders and medications were entered into a text based computerized record, I would enter the orders onto a paper order sheet.  From there the pharmacist would either write up a parallel record for what the pharmacy needed to do or enter it into computerized pharmacy software.  An MAR (medication administration record) would be used by nursing staff to record the administration and time of administration of every medication.  There was a set of checks and balances because every dose of medication was checked at some point by a physician, a nurse, and a pharmacist.  In the 1980s and 1990s, clinical pharmacists would often have close relationships with the inpatient nursing and medical staff.  Those relationships were instrumental when it came to dosage changes, using novel medications, and making sure that all of the medication was given as scheduled.  The entire chain of events in the case of a low dose gabapentin prescription would start with a very simple handwritten order like the one below:





That is all written in my notoriously bad handwriting but I think it is perfectly legible.  I wrote it to show in two places that the capsules used here were all 300 mg and how they can be increased over three days.  More importantly, I turned on a stopwatch just as I started to write this order and it took me 1 minute and 50 seconds.

Compare a recent effort using an EHR.  The scratching in red at the top of this post is basically a worksheet on how to enter the medication without making a mistake.  The overall titration is the same (except the starting dose is 300 mg three times a day), but there are large differences.  In this case the physician is responsible for entering the medication into the pharmacy record and MAR at the same time.  The convenience with which that can be done is software dependent.  With the available software there are only two possibilities - add a new line of gabapentin doses to the HS, AM and Noon doses respectively over three days or rewrite the adjacent blocks of gabapentin doses and ultimately the 600 mg TID dose.  The difference is that the first procedure involved three steps and the second procedure four steps.  Each step also involves writing in the "Comments" section on each order to make sure that there is no confusion and that multiple doses of gabapentin do not end up being given over the course of the day.  For example in the red diagram for the single gabapentin 300 mg dose at the bottom of the column on the 28th I might enter:  "This is a single gabapentin 300 mg dose in the AM on 7/28/2016.  It is a one time dose".  Using any standard EHR will generate four or five separate orders for these simple titrations.  My first time through using the top method took me 30 minutes and at the end I had broken into a cold sweats.  I had to double check all of the text orders against my sketch (boxes and U-shaped checks) and the MAR.  I ended up calling the pharmacist and giving him a verbal version of my sketch as a back up.   The second method took me a total of 15 minutes.

This very basic example illustrates some huge problems with the EHR:

1.  Fewer people have hands on the medication orders - There may of may not be an immediate double check by the pharmacist.  Nursing staff are no longer entering the MAR and double checking how it looks.  The entire task and all of the associated time has been shifted to the physician.  When this happened, clinical pharmacists also disappeared from the floors.  The hype was that we have a newer and safer systems.  It should be apparent from my example that more can go wrong with the EHR titration than more traditional methods, even if there is a clinical pharmacist at the other end reading and approving hundreds of these order entries.

2.  More and more time is added to the physician - The EHR is a classic example of how numerous jobs including billing and coding, transcribing, and now data entry that used to be done by a pharmacist has been added to the physician's burden over the years.  It is as if physicians have unlimited time for all of these additional tasks.  The time constraint has to increase the likelihood of errors in the EHR.  If you have 5 or 10 minutes between patients and have to add even a mildly complicated order - it can easily take up twice that amount of time.  Administrators view this as a plus, because other jobs can be eliminated and physicians never get paid for administrative time.  By now it should be apparent that the enterprise wide EHR is such a financial burden on organizations that jobs need to be eliminated to pay for it - often many more jobs than the physician workload has assumed.

3.  The software itself has 20th century sophistication but without the report writing capabilities or data analysis - anyone who used spreadsheet or database software in the 1990s is used to the intensive data entry approach used in the modern EHR.  Should an extremely expensive, federally mandated piece of software be this clunky to use?  Should it take me 5-10 times as long to enter an order with this software as it did by writing it down on a piece of paper?  Should the final report of a hospitalization be a phone book sized incoherent document with very little information density?  I don't think that any of these constraints should apply.  I did not include the time it takes in direct conversations with a pharmacist to clarify what was entered in the EHR.  Every home computing environment these days is at least partially object/icon based to minimize typing where possible.  In the case of medication entry, the obvious solution would allow the physician to point and click medications on the MAR with no typing.  Select the medication and dose and enter it directly into the MAR with a few mouse clicks.  That would easily beat my time for writing it out and it would be more accurate than either of the approaches that I wrote about here.

I can only speculate about all of the business and political incentives in place that has resulted in the current EHR environment.  A lot of them have been clearly documented on the Health Care Renewal blog by searching EHR.   That search will also reveal a number of safety concerns and the inescapable political factors that currently exist in a healthcare environment that routinely ignores the concerns of physicians in favor of those with no medical expertise.    



George Dawson, MD, DFAPA      


Reference:

1:  Mason BJ, Quello S, Goodell V, Shadan F, Kyle M, Begovic A. Gabapentin treatment for alcohol dependence: a randomized clinical trial. JAMA Internal Med. Published online November 4, 2013. doi: 10.1001/jamainternmed.2013.11950.

2:  Brett Boese.  Mayo Clinic tries to avoid physician burnout.  Rochester Post-Bulletin.  July 29, 2016.  Link. 

Timely article on Mayo Clinic concerns about burnout and the EHR.  The Mayo Clinic is currently in the process of conversion to Epic EHR and will "go live" on various dates between the summer of 2017 and fall of 2018.  Tait Shanafelt was interviewed about a study he co-authored on the EHR showing the clerical burden led to decreased job satisfaction and burnout.  Responding to a number of strategies to reduce physicians clerical burden his conclusion was: "The specific strategy probably used likely matters less than recognizing that physicians should not be doing this and finding a practical way to have this task completed by support staff."

3:   Shanafelt TD, Dyrbye LN, Sinsky C, Hasan O, Satele D, Sloan J, West CP.  Relationship Between Clerical Burden and Characteristics of the Electronic Environment With Physician Burnout and Professional Satisfaction. Mayo Clin Proc. 2016 Jul;91(7):836-48. doi: 10.1016/j.mayocp.2016.05.007. Epub 2016 Jun 27. PubMed PMID: 27313121.

4:  Shanafelt TD, Hasan O, Dyrbye LN, Sinsky C, Satele D, Sloan J, West CP.Changes in Burnout and Satisfaction With Work-Life Balance in Physicians and the General US Working Population Between 2011 and 2014. Mayo Clin Proc. 2015 Dec;90(12):1600-13. doi: 10.1016/j.mayocp.2015.08.023. Erratum in: Mayo Clin Proc. 2016 Feb;91(2):276. PubMed PMID: 26653297.