A State-of-the-Art Update
on Point-of-Care Testing
Gerald J. Kost, MD, PhD
Copyright © 2000 - 2001
Knowledge Optimization®. All Rights Reserved.
Part 1
Slide 1, Slide 2
This presentation includes two parts, Part I: Improving Medical Outcomes and Part II: Preventing Medical Errors in Point-of-Care Testing. The first part includes a workshop intended to identify targets for optimization of diagnostic therapeutic processes used in point-of-care testing in your own institutions. The second part is a work in progress reporting to you what United States experts feel are the needs for security, validation, performance, safe guards, and connectivity of point-of-care testing.
For additional details, please also refer to the Web site for our Point-of-care Testing Center for Teaching and Research and to the contributed book, Principles and Practices of Point-of-care Testing that I am editing for publication in 2001.
Slide 3
The goal of point-of-care testing is to improve medical and economic outcomes.
Slide 4
The definition of point-of-care testing is testing at or near the site of patient care, wherever that medical care is needed. It is important to point out that this definition does not depend on the type of instrument, per se, that is whether it is transportable, portable, or handheld.
Slide 5
The next slide lists advantages and disadvantages of point-of-care testing. Generally, my talk will follow the outline of the list under advantages. However, for every advantage, there tends to be an offsetting disadvantage, mainly due to the state-of-the-art of point-of-care instruments at this time. For example, we have patient and problem focused test clusters, which work well, as I will show you soon, but on the other hand test menus tend to be rather limited. They tend not necessarily to address the clinical problems and the difficulty that the physician faces in managing and treating patients with these clinical problems.
I will present evidence that shows clearly that point-of-care testing can decrease iatrogenic blood loss. Then, on the other hand, because point-of-care testing uses for the most part whole-blood analysis, there tends to be problems with analytical performance. Similarly, point-of-care testing provides faster therapeutic turnaround time but because of where the testing is performed and other factors, there may be inadequate quality control (QC), proficiency testing (PT), and documentation of test results. Point-of-care testing improves clinical decision making and medical outcomes but on the other hand there is the potential for unauthorized testing and increased costs of diagnostic testing per se.
Slide 6
The next slide shows how point-of-care testing improves turnaround time. The graphic on the left shows you turnaround time for pH, blood gases, electrolytes, hematocrit, and glucose in three sites: the central lab, satellite lab, and at the point-of-care. These data was published by Salem et al. in JAMA in 1991. You can see that point-of-care testing clearly improved turnaround time. But, importantly, it also decreased the blood volume required from the patient.
Slide 7
The next slide defines the concept of therapeutic turnaround time, abbreviated TTAT. Therapeutic turnaround time, as shown at the top, includes time intervals from test ordering to results receipt and result receipt to actual treatment by the clinical team. Either whole-blood analysis or point-of-care testing can decrease therapeutic turnaround time, as shown by the bars along the bottom left. Additionally, these efficiencies tend to decrease the interval, shown in green, from treatment to the desired outcome.
Slide 8
The next slide summarizes data published by Killgore et al. in Clinical Chemistry in 1998. These data address therapeutic turnaround time, shown on the left. You can see that point-of-care testing decreased therapeutic turnaround time. This was true of both a near-patient satellite laboratory in the CICU and testing directly at the bedside. The increment of time above approximately 7 minutes shown for bedside testing were the increments of time spent by clinicians in considering the electrolyte and other laboratory test results. The middle graphic shows that satisfaction was higher with point-of-care testing. The orange refers to accuracy, the white to overall satisfaction. The right bars show the evaluation correlates resulting in the higher satisfaction, namely that point-of-care testing improved patient care, conserved labor was more convenient and was also more timely. On the other hand, in this study, clinicians were not quite as concerned about the accuracy in terms of evaluation correlates.
Slide 9
Now, let us turn to the critical care profile and concept of test cluster. The next slide shows vital functions and their diagnostic pivots. From a physician's prospective, it is very important that test clusters offered on a point-of-care instrument cover the vital functions you see on the left. For example, for acid based management, management of ventilation, management of oxygenation and energy, glucose, hemoglobin, PO2, O2 saturation, pH, PCO2, total carbon dioxide content, and bicarbonate are extremely important. For the management of homeostasis problems, you can see that there is a similar list of critical tests. These are the short lists. These are the lists that one should look for in a point-of-care instrument in order to make sure that the instrument and its test clusters or test cluster will be clinically effective and efficacious. Let's take a specific example.
Slide 10
The next slide shows a test cluster which has currently been shown to be quite cost effective in the evaluation of patients with chest pain in the emergency room setting. We have implemented an algorithm that uses cardiac troponin I (cTnI) myoglobin as the two primary tests to identify myocardial injury with secondary CK-MB mass testing on an as-needed basis for event dating. We will discuss the use of algorithms later.
Slide 11
The advantage of this test cluster is it allows you to either rule in or rule out myocardial infarction in patients presenting with chest paint and it also has other advantages such as dating of the time of the actual event. As you know, rapid treatment of coronary occlusion is paramount in the case of a patient presenting with acute myocardial infarction and, therefore, the reduction of therapeutic turnaround time, that is the time from ordering to actually treating a patient, is absolutely essential in conjunction with this particular test cluster for the evaluation of patients with chest pain.
Slide 12
A patient presenting to emergency room or a patient in critical care manifests an outcome. Let's now turn our attention to, in the next slide, how outcomes integrate preceding effects. We know that, in fact, this is the case.
Slide 13
Therefore, we must integrate our different strategies to improve patient outcomes. The following series of slides starting first with integrative strategies will focus on and show you evidence for how this can be done.
Slide 14
The next slide summarizes an integrative strategy, published by Steffes et al. in Clinical Chemistry in 1996. This strategy uses point-of-care testing. The problem being addressed was utilization of diagnostic testing in a surgical ICU and a tertiary care university hospital. The goal was to improve use of resources. The study design was: There was an on site laboratory. Whole-blood analysis was used. An ordering protocol was implemented. Required daily renewal. There were temporal test sequences, for example, for diabetes. Work flow was optimized. They used patient focused test clusters and other features of testing were reengineered for efficiency. In summary, the results showed faster turnaround time and blood conservation with the use of point-of-care testing. And the authors felt that point-of-care testing in the integrated strategy were effective.
Slide 15
The next slide summarizes their results graphically. As you can see, either with routine or emergency cases (on the left) turnaround time was decreased with the integrated protocol and point-of-care testing shown by the orange bars. Confirming the results of the other study that I mentioned, Steffes found that sample volume was also decreased and the blood removed from the patient in milliliters per 24 hours was decreased.
Slide 16
Let's turn to the concept of use of the algorithm.
Slide 17
Despotis published a paper in the Journal of Thoracic and Cardiovascular Surgery in 1994 with the results summarized in this slide titled Integration Facilitates Optimization. The problem: cardiac surgery with cardiac pulmonary bypass and microvascular bleeding. The goal of their team: to avoid unnecessary transfusions and empirical treatment. The study design was a comparison of onsite hemostasis testing with algorithmic treatment versus hospital laboratory testing and standard therapy using a conventional approach. The results: They showed that the rate of blood loss and transfusions decreased. That turnaround time and operative time were decreased. And though not statistically significant the number of explorations for postoperative bleeding also decreased. Their conclusion: Onsite hemostasis testing guides specific therapy and optimizes the treatment of microvascular bleeding due to platelet disorders and factor deficiencies.
Slide 18
The next slide shows the algorithm that they used where point-of-care testing included onsite prothrombin time (PT) and activated partial thromboplastin time (aPTT), as well as platelet count and consideration of these in a routine that without going into detail basically sought to reduce the use of expensive and potentially hazardous blood products.
Slide 19
The next slide summarizes their results graphically. You can see (on the left) that the use of onsite hemostasis testing with algorithmic treatment shown in yellow, decreased units transfused uniformly. In the center graphic, the same approach, decrease estimated blood loss and the turnaround time and operative time were also decreased, shown on the right.
Slide 20
We have discussed integrated strategies and algorithms.
Slide 21
Now, let's turn our attention to performance maps. The next slide shows the conventional flow of information for the evaluation of a critically low sodium. Critically low sodium can result in encephalopathy, seizures, or respiratory failure. This occurs around the value of 120mmol/L. Unfortunately, laboratories do not agree as to the actual critical value for a low sodium. This can result in indecision. If a low sodium is discovered, it still has to be decided whether it is below the critical level. Following the green blocks then, if it is below the defined critical limit, it has to be successfully notified on an emergency basis to the physician. In fact, once that is done, it still could be a result reflecting pseudohyponatremia, a condition that results from hyperlipidemia or hyperproteinemia. Direct whole-blood analysis can be used to rule out pseudohyponatremia, but in fact a repeated or duplicative technology using whole-blood analysis could result in questions as to interpretation and certainly will result in time delays.
Slide 22
Now, let's consider the performance map approach to this problem as shown in the next slide. We have used this approach for the evaluation of low sodium for nearly two decades in our operating room near-patient laboratory. Point-of-care whole-blood analysis can lead to rapid and accurate evaluation of a suspected low sodium. Because it is a direct measurement, whole-blood analysis does not encounter the problem of pseudohyponatremia. The low sodium reported directly to the clinical team can result in the rapid design of a repletion plan at a rate that is appropriate for the patient. The treatment then can pivot on its speed and whether the threshold of the sodium has been achieved. The evidence trend can subsequently be followed in this feedback loop so that the critical level is eliminated. This performance map has the benefits of eliminating time delays and delays in processes or additional processes that can delay the accurate diagnosis and treatment of critical hyponatremia.
Slide 23
The next slide shows diagnostic-therapeutic process information for ionized hypocalcemia. You can see in the bars as you start at the top and move downward that they move to the left resulting, if the ionized calcium is low enough, in cardiac arrest. This is information that, in part, we obtained during our initial encounters with liver transplantation where citrate from transfused blood combined ionized calcium and takes it out of the active pool that the heart needs. You can see, and this is the result of empirical observation in the 1980s, that the treatment threshold corresponds well to critical limits which were established in another national survey that I did. What the question is then from the physician's perspective, how can a problem like this be successfully attacked and how can we deal with it efficiently.
Slide 24
The next slide suggests a performance map for the detection and monitoring of ionized hypocalcemia. Rather than measuring total calcium after considering an abnormal level, we go directly to the point-of-care where whole-blood analysis yields an ionized calcium measurement. We can decide with the clinical team whether this is a critical result and, if necessary, treat the patient immediately. The advantage is with cluster analysis provided by other tests provided on the point-of-care instrument, we are able to consider the whole picture of the patients clinical symptoms and signs and dose and time our treatment according to the evidence trend in this feedback loop. The overall advantage is being once again that we have optimized the diagnostic-therapeutic process and done this is in a timely fashion.
Slide 25
The next slide summarizes the goal of the performance map approach. In fact, this should be the same goal of selection of point-of-care instruments and their implementation for clinical effectiveness and efficacy from the physician standpoint. In picking the examples that we discussed specifically low sodium and low ionized calcium, we were addressing surrogates of adverse outcomes. These particular analytes reflect at the point-of-care key pointers which we must detect, correct, prevent, and optimize in regard to pathophysiological events that may occur. These surrogates of adverse outcomes feed forward to sub-lethal events which result in morbidity and lethal events which result in mortality and ultimately reflect themselves in the integrated medical and economic outcomes of our patients. Please remember that outcomes integrate preceding effects, therefore, integrate to improve outcomes.
To summarize the presentation to this point then -- we have addressed therapeutic turnaround time. We have shown that point-of-care testing decreases therapeutic turnaround time and that at the same time saves patient blood volume. We have worked with a critical care profile to identify certain test clusters which are highly effective and in the most recent series of slides, we have shown how integrative strategies, algorithms, and performance maps can improve patient outcomes while addressing key clinical concerns of the physician.
Slide 26
In the next slide let's turn to the new critical care mosaic. This side portrays modern critical care testing in the ICU as a mosaic of in vivo, ex vivo, and bedside modalities. (This figure was provided courtesy of Dr. Neil Halpern.) I want to just briefly show you examples of ex vivo and in vivo modalities.
Slide 27
The next slide shows a device which samples from the radial artery and returns blood after performing laboratory testing. The measurements are shown along the top. This again is a critical care profile.
Slide 28
The next slide shows an example of an in vivo sensor set inserted in the patient's radial artery.
Slide 29
In the next slide, we received that these advances will be defined in terms of a new mosaic that includes in vitro, in vivo, ex vivo testing. Once again, integration is the key to effectiveness and efficiency because these modalities will bring together both biochemical and physiological monitoring. Ultimate benefits that we should look forward to are reduced costs and improved outcomes.
Slide 30
In the next slide we can see some of the difficulty, though, in relating this new mosaic and its use of point-of-care testing to specific diagnosis-related groups (DRGs). The challenge of managing an operating room or intensive care unit setting, or for that matter an emergency department, is that there is a heterogeneous population of patients and diagnoses present represented by the DRGs one through five on the right axis. In fact, we have to efficiently step our patient through the diagnostic-therapeutic processes that correspond to each of these DRGs. It is not easy to optimize an individual patient pathway shown by roman numeral I and additionally studies that look at the overall effects of introduction of point-of-care testing, for example, can be misled if they look broadly across all of these DRGs at once, as shown by roman numeral number II.
Slide 31
The next slide presents to you the physician's prospective on this and a probable method by which it will be solved within the next few years. This slide, without going into detail, is a feedback system for outcomes optimization. A typical system, which is currently in use, is called Project Impact, invented, promoted, and managed by the Society of Critical Care Medicine in conjunction with Tri-Analytics, a database expert company. You can see in this flow chart that the patient outcome data set feed into the critical care knowledge base using various metrics, such as morbidity, mortality, function, severity of illness, resource utilization, and cost. These then feed into either the physician outcomes optimizer or the national local data repository for review and evidence-based patient management. We complete the cycle on the top and the bottom, most importantly, with performance and quality enhancement, which then is a feedback mechanism to improve our critical care knowledge base.
Slide 32
The next slide translates this into down-to-earth terms and shows the potential pitfalls that can stymie outcomes optimization. Along the top row are represented those things that relate to temporal optimization and along the bottom row those things that relate to diagnostic-therapeutic process optimization. From the physician's prospective, laboratory testing, satellite laboratory testing, or testing at the point-of-care can be cumbersome, if not downright harmful to the patient if an adequate number of tests are not performed simultaneously, if testing is not appropriately integrated with the clinical care path or the treatment routine that is necessary, or if results are not communicated successfully to the clinical team. On the other hand, the diagnosis and treatment can be put at jeopardy if test results are not accurate, the test cluster is not complete, and hence observations as well as diagnosis could be missed or delayed, and finally, if the test cluster and process of diagnostic testing is not properly focused on the particular problem which is of concern to the physician.
Slide 33
The next slide shows a mechanism by which we can relate these potential pitfalls directly to point-of-care testing and possibly avoid some of them. This presents the concept of a multitasking interface for point-of-care testing as a work station. This product is currently offered by one vendor and in a way represents the analog to the physician's attempt, in the outcomes optimization flow chart I showed you earlier, to solve the problem integrating and collecting information which simultaneously relates both to several patients and their DRGs. In the multitasking interface shown here, chemistry, glucose, blood gas, manual tests, hematology, and coagulation are integrated, potentially reviewed by a designated technologist, who either accepts or rejects the data, and then forwards the results to the appropriate computerized information system. Lack of complete connectivity for point-of-care testing is probably the biggest problem currently faced by the field at this time. Conversely, when connectivity is properly provided, and this should be a key prerequisite for instrument selection from the physician's standpoint, I think we will see a new trend towards increased use of point-of-care testing and then the added advantages of being able to relate the results in the database generated at the point-of-care to other programs for outcomes optimization such as the one that I just showed you.
Slide 34
The next slides concludes my presentation (Part 1). Point-of-care testing is the standard of care for rapid response diagnosis and life threatening crisis in emergency resuscitations. Physicians really do expect a turnaround time that is a therapeutic turnaround time of five minutes or less in these situations. Rapid test results facilitate immediate therapeutic decisions and interventions in critical care and we saw several examples of how point-of-care testing can be integrated to successfully achieve this. In vitro, ex vivo, and in vivo technologies enable simultaneously monitoring of biochemical and physiologic variables. This is certainly the next wave of point-of-care testing as connectivity through, for example, multitasking, merges the knowledge base, optimizes knowledge essentially for the physician at the bedside. Finally, knowledge integrating systems are essential for modern evidence based medicine and outcomes improvement. I showed you in the slides of performance maps what we really must target with very effective clusters and the surrogates of adverse outcomes if we wish point-of-care testing to be maximally efficacious.
Slide 35
Now, let's conclude Part 1 of the presentation with analysis of DRG data and two summary slides as follows: The first slide summarizes key observations. Notable medical problems require integrated diagnostic-therapeutic strategies. Complications increase morbidities, increase costs, and length of stay. Increase is severity of medical problems increases incremental daily costs. Reductions in costs and length of stay are essential to match reimbursement. Laboratory efficiencies directed at reducing length of stay are crucial.
Slide 36
The second slide addresses optimization. Target problem, prevention, length of stay, and complications are co-morbidities. Synchronize outcomes management to avoid increases in severity. If you look at the data you will see how increases in severity are related to striking increases in charges. You will also see that laboratory testing represents about one day in terms of charges for a given DRG. Therefore, laboratory strategies for diagnostic testing, as well as point-of-care testing strategies, should be designed toward decreasing the severity of the condition very early on. Therapeutic turnaround time and other factors that we have addressed are essential in this process. Design point-of-care testing to be simultaneous, integrated, and communicated. Use accurate, complete, and focused test clusters. Lastly, incorporate connectivity solutions to accelerate optimization of diagnostic therapeutic processes.
(Click here to download the DRG data tables)
Handout#1 Handout#2 Handout#3 Handout#4
Now, please refer to the four DRG data tables. Use these data to identify targets for optimization of diagnostic-therapeutic processes using point-of-care testing in your own institutions based on the information in the two summary slides I just presented for this workshop portion of Part 1. This concludes Part 1.