Healthcare Economics  

Determining Value in Healthcare
The final installment in the five part series and the presentation of a new model for evaluating healthcare services

Part 5 of the 5-part series

Determining Value in Healthcare The complete series can be found online at

Summary of Previous Articles

The first article in the series emphasized the difficulty in determining “value” in healthcare but nevertheless established a definition of “value”. Here is the difficulty - most people would agree that a 21 year old male who has appendicitis, presents to the emergency room, has his appendix removed uneventfully, returns home in one day, goes back to college or work in seven days, and then returns to a normal life certainly represents “value” for healthcare resources spent (approximately $15,000 spent but the result is a long, productive life). On the other hand consider this example – a 78-year-old female who has a history of high blood pressure, diabetes, and heart disease, presents with chest pain, is admitted to the intensive care unit, stays there for eight weeks, and then passes away (approximately $400,000 spent). Was there “value” in spending $400,000 in eight weeks? When healthcare is involved, and consequently people and emotions, it is very difficult to establish “value”. It is difficult to satisfy every stakeholder’s definition of “value” – however a decision has to be made, and “value” from society’s perspective seems to be the best way to approach the concept of “value”. The subsequent four articles provided a historical perspective on the evolution of healthcare in the United States and described how disparate, but powerful, stakeholders molded the system to serve their individual interests. This led to a fragmented, expensive healthcare system that, as of 2010, excluded 50 million Americans from access to healthcare. These articles also highlighted critical lessons learnt by analyzing the historical development of the US healthcare system and provided counsel to countries in the Middle East that are in the process of developing their healthcare systems.


Providing comprehensive healthcare to all citizens and creating “value” at the same time is a challenging task. To create “value”, healthcare systems need to measure, quantitatively, elements critical to the efficient and cost effective delivery of high quality, universal health care[1-4]. Perhaps more importantly, these critical elements need to be measured and presented in a way that allows easy visualization of how well the health care system is functioning and where it needs improvement. How well healthcare is being delivered, to what percentage of the population, and at what cost (i.e. “value”) are some, but by no means all, of these critical elements. Historically, and as described in the previous articles, the fee-for-service model in the United States fostered the development of a health care system that measured and improved financial returns and paid little, if any, attention to outcomes, prevention, or efficiency. The more the hospitals and physicians did, and the more the cost, the more they got paid. As Paul Starr, the renowned medical historian, has pointed out, “…a better way to exponentially inflate the cost of healthcare could not have been devised”[ 5]. Thus, if we are to demand “value” then we need to radically change the way we think about healthcare and begin to measure parameters relevant to optimal and “value” driven delivery of healthcare. In other words, we need a self-learning healthcare system that has as its premise high quality care, continuously improving efficiency and “value”, and comprehensive healthcare delivery to all citizens. In this article such a model of measuring the healthcare system is proposed.

The model – The Triangles

The proposed method of evaluating our healthcare services should be comprehensive enough that it measures all the relevant parameters, be flexible enough to change – and change fast, and be able to give an overall view of how efficiently healthcare is being delivered. A simplified version of the basics of such a model is provided in Figure 1 (please read figure legend before proceeding). It would be critical for Middle Eastern countries that are in the process of developing their healthcare systems to keep such a model in mind as it would help avoid the costly mistakes made by the US over the last 80 years.

Figure 1 is an abbreviated and idealized introduction to this model. In this model, line BD (Birth, Death) represents the continuum of life without disease or, at least without disease that prevents an individual from performing the daily activities of life – i.e. baseline, productive functioning. Prevention of disease is, by far, the most cost-effective medical intervention and the proposed method of measuring healthcare delivery will provide, at a glance, the percent population that is utilizing adequate preventive healthcare. For example, in Figure 1, only 47% of the population (all numbers in all figures are percent of total population) uses routine primary care visits (RPCV). Additionally, none of the inadequately insured individuals is using RPCV. This provides clear indication of where improvement efforts need to be expended. If an individual falls off the productive continuum of life (that is, falls off line BD), as everyone does at some point in time, he/she enters one of the triangles of the healthcare delivery process. Here, the goal should be to treat the individual as quickly, efficiently, and cost-effectively as possible and return him/her to line BD (i.e. baseline functioning). At a glance, this representation of the healthcare delivery process also tells us what percent of the population is inadequately insured. Most importantly, regularly looking at this representation and seeing how many individuals have inadequate insurance, is a constant and necessary reminder that consistent efforts should be expended to reduce the number of inadequately insured (red triangles) to zero (many European countries have achieved that). These are some examples of how this model might be used.

Figure 2 demonstrates a more realistic version of what such a model might look like (please read figure legend before proceeding). As noted above, the triangles each represent a particular service in the health care delivery system – such as routine primary care visits (RPCV), outpatient procedures (OPP), etc. Note that line BD, and the ‘inflow’, ‘process’, and ‘recovery’ sides of each triangle, are designed in a way that emphasizes the premise that prevention, that is staying on the black line BD, is the best option and that, if an individual “falls” into a triangle, then swift return to line BD (baseline functioning) is critical. Thus, the model keeps at the fore, that if entry into a triangle is necessary, it is best to ‘optimize’ (1) the ease of getting there (inflow), (2) the efficient delivery of the treatment (process), and (3) the rapidity with which the individual is returned to line BD and once again becomes a useful member of society (recovery). How then do we optimize inflow, process, and recovery?

Figure 3 demonstrates the details that underlie inflow, process, and recovery, and how “value” can be extracted by using these triangles to document the health care process (please read legend before proceeding). Each side of every triangle, or an independent healthcare service, needs to be measured in multiple ways. Ideally, every aspect of inflow, process, and recovery needs to be measured. However, that may not be necessary. It may be that only certain critical elements are sufficient to adequately measure and monitor the ‘inflow’. Similar measurement parameters can be created for ‘process’ and ‘recovery’. How many, and which parameters should be measured, will vary considerably for most triangles. That is the inherent flexibility in this system. It may be that the inflow process for RPCV needs only 3 measurements to determine if it is performing well whereas the inflow process for OPP may need 17 measurements. However, what needs thought and analysis is what are the critical elements that need measurement, how will they be measured and analyzed, and how will they convey the level of function and efficiency of any given healthcare service. If parameters are identified that are critical, but cannot be measured, then methods to measure them should be developed. An additional flexibility in this model is that any number of triangles, or healthcare processes, can be added into this system (see Figure 4). Thus, the model will accommodate growth in healthcare services. The more triangles that are added, the more comprehensive the model becomes – and this leads to more actionable data that provides precise areas to target for process improvement.

It is probably evident by now that underlying inflow, process, and recovery, or each side of any triangle, is a lot of data regarding the critical elements that measure the efficiency of that side. How should this data be analyzed? This can be done in two ways and displayed as a “dashboard” (not shown in figures). Whereas a complete discussion of how the underlying data can be analyzed and a dashboard created is beyond the scope of this article, a brief summary would be useful. The data can, in general, be analyzed in two ways. First, one can compare a certain critical element to those of many other hospitals and determine a percentile rank. Second, one can decide on an absolute goal. Let’s look at an example to clarify the concept. Let’s take OPP which is utilized by 30% of the adequately insured population and look at one critical element for process shown in Figure 3 – what percentage of first operations start on time? One could use a percentile ranking and determine that in a particular hospital 65% of the cases start on time - which may represent a percentile ranking of 78 (that is, if compared to all other hospitals, this particular hospital is better than 78% of the hospitals looked at). Thus, this hospital, whereas reasonably efficient compared to other hospitals, can still improve. On the other hand it could be determined that if 90% of first cases start on time, that represents an efficient process, thus creating an absolute measure against which all hospitals would be judged. Either one or the other is acceptable depending on the critical element measured. Of course, cost for all sides of the triangle – inflow, process, and recovery – will be critical to measure so that “value” can then be determined. The reader will by now be imagining a very complex “dashboard” that underlies these rather innocuous looking triangles – and the reader would be correct. A complete model would measure hundreds (if not thousands eventually) of parameters, analyze them in a statistically sound manner, and determine a relative, or absolute, numerical ranking for that parameter. So, e.g., an analysis might reveal that ‘inflow’ to OPP can be improved in a certain hospital by decreasing wait times since their wait times are much longer than a previously determined, acceptable or optimal, duration. A similar analysis would also be made for all the measured parameters for all sides of all triangles in the model and a final “dashboard” would reveal where improvement efforts are needed. Thus, the overall model, shown in Figure 2, would identify a general area (inflow, process, recovery, RPCV, etc.) that needs improvement. The dashboard would then provide the details of exactly what critical elements within that general area need improvement. If the overall model shows there no critical elements present then these need to be defined and measured.

Figure 4 demonstrates what the final “bird’s eye view” of the model of measuring healthcare delivery might look like (please read figure legend before proceeding). It is important to note that the asterisk (*) in this figure only states that the parameters, or critical elements, needed to determine the efficiency of that particular side of the triangle have been determined – it does NOT state that these parameters have been optimized. Optimization can only be determined by looking at the dashboard discussed above (not illustrated). In analyzing Figure 4, one can arrive at the following major conclusions: It can be used for any given population – enrollees in a small insurance company, a city, a state, or an entire country (depending on how accurately and uniformly the data can be measured). 1. Only 50% of the population is using RPCV. Thus adequate prevention is deficient 2. None of the inadequately insured are using routine primary care visits (RPCV) 3. 23% of the total population is inadequately insured and 77% adequately insured. 4. Many healthcare services, or triangles, have sides that don’t have definitive, accurately measured, parameters that could evaluate adequate and optimal functioning 5. 10% of the population does not return to line BD (baseline, productive functioning) 6. The inpatient procedures and admissions (IPA) triangle do not have adequate critical elements defined that measure this service’s efficiency or “value”. 7. More triangles are needed and can be added whenever necessary 8. Of the triangles that have asterisks (*) it would be necessary to evaluate the dashboard to determine if optimal function has been achieved for either inflow, process, or recovery 9. This diagram could be compared to last year’s to determine if there have been improvements in certain triangles – e.g. has the RPCV triangle gotten bigger or smaller or, in other words, is preventive care reaching an increasing percentage of the population?

The above are just some of the conclusions evident from Figure 4 – there are many more. This model, representing the functioning of the healthcare system, provides actionable data which guides healthcare leaders to direct resources towards the most critical deficiencies.

Conclusion and Summary

A complex system such as healthcare delivery cannot be completely described and evaluated in a short article. However, what can be done is provide the overall goal, direction, and the impetus to change the way we think about and evaluate healthcare. The model presented in this article attempts to do just that – provide an overall direction and a way to measure whether we are indeed moving in that direction. In this model, the overall goal is to create a self-learning healthcare delivery system that continuously improves “value” delivered while providing comprehensive, universal healthcare.

There are myriad authors that have opined broadly on “value” in healthcare (please see first article[6]) and its references) and the entire range of issues are beyond the scope of this article. However two points merit repetition. First, “value” as defined in the first article[6] and as is usually, and broadly, defined[3, 7] – ( outcomes/cost) – is a ratio, and one could argue that inferior care provided at even lower cost could increase “value”. That is, of course, a valid criticism[8] and it should be emphasized that “value” as defined here is only, well, valuable, if the outcomes remain the same or continue to improve. Second, the question of “value from whose perspective” is important[8-10]. Of the many stakeholders identified (society/ economic, patient, provider, payer, employer, health product/device manufacturer), it would seem at first that it would be hard to reconcile any two stakeholders’ perspectives. However, if any two perspectives should be, and perhaps even can be, reconciled, they should be those of the patient and of society/economic. The argument would go as follows: if the healthcare system can be optimized from an operational standpoint and made efficient enough then there should be enough money left to take care of any individual patient’s unique needs. The method and model provided in this article has the potential to reconcile these two perspectives. Additionally, this model proposes to measure the correct parameters, or critical elements, in the correct way[2] and apply that knowledge to create a self-learning healthcare system that will provide efficient, high quality healthcare for the optimum cost.

This model is clearly a work in progress and will take decades to perfect. However, the journey must begin – and difficult questions like what is “adequate insurance”, what is “optimal functioning or efficiency”, what elements are indeed “critical” to measure, and myriad others will need answers. As long as we head in the right direction, the complex task of perfecting a model that effectively evaluates the healthcare system will no doubt be completed.


1. Khan, A., Thirty-day Readmission Rates as a Measure of Quality: Causes of Readmission After Orthopedic Surgeries and Accuracy of Administrative Data - Practitioner Application. Journal of HealthCare Management, 2013. 58(1). 2. Kaplan, R.S. and M.E. Porter, How to Solve the Cost Crisis in Health Care. Harvard Business Review, 2011(September). 3. Porter, M.E., What is Value in Health Care? New England Journal of Medicine, 2010. 363(26): p. 2477-2481. 4. Porter, M.E. and E.O. Teisberg, Redefining Health Care 2006, Boston: Harvard Business Review Press. 5. Starr, P., The Social Transformation of American Medicine 1982: Basic Books. 6. Khan, A., The Importance of determining ‘Value’ in Healthcare. Middle East Health, 2012(Jan-Feb). 7. Young, P.L., L. Olsen, and M. Mc- Ginnis, Value in health Care - Accounting for Cost, Quality, Safety, Outcomes, and Innovation - Workshop Summary, in The Learning Healthcare System Series, I.O. Medicine, Editor 2010, The National Academies Press. 8. Tilburt, J.C., V.C. Montori, and N.D. Shah. Letter to the Editor. New England Journal of Medicine 2011; Available from: 9. Stuart, B. and B.A. Chernof. Letter to the Editor. New England Journal of Medicine 2011; March 11, 2011:[10.1056/ NEJMc1101108]. Available from: 10. Cohen, A.J. New England Journal of Medicine 2011; March 11, 2011:[Available from:]    

Arby Khan, MD, FACS, MBA was the Deputy National Director for Surgery for the United States Veterans Health Administration - which oversees 151 hospitals and more than 1000 outpatient clinics. Dr Khan is a regular contributor to Middle East Health. He has written on a range of subjects – such as Human Resources management in hospitals, Change Management in GCC hospitals, Brain Death and Hospital Resource Management and organ transplant- related legislation, among others – with a view to improving healthcare in the UAE and the wider region. He is a multi-organ Transplant Surgeon and Immunologist and has successfully started, from the ground up, two multiorgan transplantation programmes – one in the United States and one in Abu Dhabi. He is the author of many clinical and basic immunology papers, and has been educated, trained and employed variously at University of California - Berkeley, McGill University, University of California - San Francisco, Harvard Medical School, Yale University - Graduate School of Immunobiology, University of Pittsburgh - Starzl Transplantation Institute, University of Vermont - School of Medicine, and Columbia University (NY). He also holds an MBA, with Distinction, from London Business School. – The views expressed in this article are those of the author and do not necessarily represent the views of the institutions for which Dr Khan has worked or currently works. 

 Date of upload: 18th Jul 2013


                                               Copyright © 2013 All Rights Reserved.