Tuesday, November 26, 2013

Healthcare: the Value Stream of care

In Lean process improvement, one of the first steps you ever take is to walk the process. For manufacturing this means to go down to the floor where the product is made and walk with one piece from beginning to end. This provides the manufacturer insight into where there's a great deal of waiting for product to come, leading to idle workers, where there's a lot of inventory piled up - a bottle neck, if there's a lot of rework - fixing defects like re-etching a person's name on the back of an iPhone, and how the material flows around the floor. This works pretty well with doctor's offices too. You can draw a map to all the different places the Dr. walks, the nurses walks, and where the patient walks. Any transportation in a Lean system is waste, so reducing that is important.

Mapping value streams essentially take this to the next level. You map all the major steps that the material that goes into your product step through before and after you. This allows a manufacturer to see all the waste before and after them, enabling them to partner with their suppliers and customers to reduce waste and unnecessary processing. For example, many of us have worked retail. Some times when you do stocking you'll find shirts that are in bags that are in a box. This is non-value add because it's highly unlikely that the bag would protect the shirt from getting wet in the case of a flood. So, it's a waste of plastic for the bag, putting the bags on the shirts, and removing the bags from the shirts are all waste. Which increases the cost of a shirt. However, there's a beginning and end of the value stream likely starting with cotton and ending with the final sale to the customer. In the case of a can of cola, it takes 319 days from the mining of bauxite to the consumption of the cola with only a total of 3 hours of actual processing of the material (Lean Thinking, Womack).

Value Stream for a can of cola through bottling (Womak Lean Thinking)
Why such a long introduction? Well, the value stream for healthcare is completely different. The beginning is when you're born and the end is when you die. Otherwise, every activity you partake in impacts your health and the eventual cost of any episode of care. An episode of care is what happens when you directly interact with a provider, hospital, or health insurer. Arguably, these are the exception to your normal behavior and take you out of your normal routine.

Thinking about health in a value stream like this is non-intuitive for providers and insurers alike. As both have accounting practices and treatment plans that focus mainly on the episodes of care and minimize the remainder of the activities a member does. Thinking in this manner places more importance on preventative care, longer term plans for mental and physical rehabilitation, and care networks for long term diseases. This is a serious shift that is starting to occur in many insurance organizations, but aren't very effective yet. The most effective portion of those three are the networks of patients that have a similar disease, such as Crohn's Disease.

I believe that looking at care in this fashion will help redesign the manner that care is designed as it will focus on different portions. As my friend Rachel pointed out, behavioral health issues are typically undervalued in the value stream of healthcare. However, with this model long term care issues should be given priorities as they impact the highest percentage of the value stream. It would also force insurers and providers to look at addressing care holistically and providing the best care in the best way when they can. Sending patients to clinics that can quickly treat conditions as cheap as possible.

I'm extremely interested in how this will play out at my company as we think more holistically about value streams for health care. Checkout my last two blogs about health care:


Monday, November 25, 2013

Healthcare: Why do we need medical policies?

In my last blog post, I discussed how US health insurance companies decide what to pay for, what information they need to make decisions on paying for care, and some of the ways they go about making those decisions. I only briefly discussed why medical policies are required and why there's a lot of complexity around health care. I hope to shed some light on this and help people understand the difficulty that both Providers and Insurers have in dealing with the cost of health care.

First, not all diseases are equal. This one is pretty obvious, but I'm not even talking about how severe the diseases. That's vitally important of course, but even a "minor" disease can lead to long term impacts to quality of life because we don't know how to treat the disease. In some of these cases it might just be embarrassing, such as if I eat Gluten and I'm out and about the next day. I'll probably have some serious issues and won't be able to enjoy myself while I'm out. There are two reasons for that. One, it's not really obvious what condition I have. Two, there's no treatment to allow me to eat gluten other than "avoid" wheat, barley, and gluten. Which in many cases is rather difficult.

This brings me to the second reason why there's a lot of difficulty - it's difficult to even diagnose what disease people have. So, for gluten issues you have only a few options, one is a blood test to looking for an immune response to a gluten or to have a colonoscopy to see what sort of damage has been done to your large intestines. Because of this lack of precision, in many cases it can lead to the matching the wrong treatment for the right disease, vice versa, or wrong treatment for the wrong disease in the most extreme cases.

Precision of Diagnosis and Precision of Treatment matrix
In the above picture from "The Innovator's Prescription" it's clear which diseases are "better" to have, those in the upper right. While those in the lower left are much harder to treat and have less consistent outcomes. For anything in between the cost and quality of care is really dependent on the experiences of your provider. I believe that this is where insurance companies can add a lot of value. Using medical policies and partnering with providers they can artificially expand the experience of a provider through providing the latest scientific research and results for a treatment and disease interaction. This will help allow providers to focus on care while getting the latest medical news from their network of insurers.

This really puts the model on it's head as the provider can take advantage of the diverse networks they are part of to learn different components of research based on the focus of those providers. I think that a true partnership between insurers and providers really will drive down health costs.

This complexity is unfortunate, but is truly part of our human condition. One way to reduce costs is to increase the amount of research that pushes care into the upper right from the bottom left. Otherwise, it's difficult for an insurer to determine which providers are taking advantage of patients and which ones are honest. There's imperfect information on both sides and the patients pay for it in the long run.

Sunday, November 24, 2013

Heathcare: How an insurance company decides to pay for what

As many of you are aware, I've left the high tech sector and have moved to work for a Health Insurance company. Since my job is to analyze processes and improve them using Lean process improvement methodology, I have a pretty unique insight into the broader workings of my company - which will only grow as I expand Lean through more parts of the company. As a note, these views are my own and do not represent the views of Cambia Health in anyway. I believe that it's important to help educate the broader public about what happens when you want to have elective care or even while you've been admitted for care - either after an elective procedure or in an emergent case.

I know I'm opening a can of worms here, so I think it's important to note that these are truly caring individuals that want their members to receive the best care. However, they are also put into a tough position because health insurance is a contract and your employers negotiate types of care that will be paid for within your health insurance plan. Furthermore, as a way to control the amount of money the insurer pays they will "manage" the type of care that is being provided to a member. This of course is the part that people hate, my doctor says this should be done, so it should be paid for by insurance company! Well, yes and no.

In some cases your doctor may recommend a specific type of treatment that you actually need, but your insurance company won't pay for it! Why not? Well, there could be a few reasons. First, you may not actually have the benefit. For example many cosmetic procedures aren't covered and aren't part of your benefit package, or maybe you've already exhausted the benefit for instance number of days in a skilled nursing facility. Unfortunately, insurance premiums are based upon the amount of benefits that you could use and the total amount of risk that you are to the insurer. Which of course is difficult, because a person may need that care, but they didn't pay to have all of it fully covered. The other unfortunate portion is that most people truly don't understand what their benefits are, which makes this even more difficult. This of course goes back to Prospect Theory where insurance companies are acting as much more rational entities than their members, which leads to a definite imbalance of understanding and event potentially power in the relationship. Humans don't read all the contracts and Econs do - insurance companies are Econs extremely rational (in the economic sense) when dealing with contracts and risk.

The second case where an insurance company won't pay for a service relates to medical necessity. This is the part where providers get really upset with insurance companies. Essentially, this is a case where the insurance company is using a combination of medical research to create policies with criteria for procedures. So when you are trying to receive care, the Diagnosis is less important than the service that your provider selects for you. Diagnosis plays a larger role whenever the insurance company recommends alternative methods of care that lead to the same level of care but typically cost less - either to the member, insurance company, or both.

This requires the provider to submit clinical information to justify that procedure. Some of this requirement may actually drive up costs because the provider may have a great deal of intuitive experience with a specific type of diagnosis and knows that the best treatment is X. However, the insurance company requires that for procedure/treatment X that tests a, b, c must be run with results q, r, s. Based on the combination of the results the insurance company determines if that procedure is medically necessary or not. This works really well when there's very clear medical research and clear correlation between diagnosis and treatment. It doesn't work as well for less precise treatments such as behavioral health - which is much more trial and error and requires a lot of time.

Who creates the medical policies though? A combination of research clinicians, internal MDs, and providers - if they are insurance company specific policies. However, there are more general policies that are recognized and used for inpatient Medicare and Medicaid. These may have more input from the people using them. Policies take time to develop and typically lag the latest research. However, in many cases these policies may represent more knowledge about a specific treatment methodology than a general practitioner or even a specialist can know because of the breadth and depth of medical knowledge.

I will discuss more topics related to healthcare and how these impact costs over the next few weeks.

Saturday, November 23, 2013

Minimum Wage, labor, and economics

Last night on the twittars I had an invigorating discussion related to minimum wage, the mobility of labor, and the economics. The basics of the argument started over the fact that in MA minimum wage will be slowly going up from around $8/hour to $11/hour. $8/hour already is pretty good for the US where the federal minimum wage is closer to $7.25/hour - which is an annual salary of $15,080 (poverty line for an individual is $11,490/year). Obama is proposing to raise the national minimum wage to $10.10/hour or $21,008/year and MA's eventual minimum wage at $11,440/year. If minimum wage had kept up with productivity gains seen throughout the economy a study argues that the minimum wage should actually be $22/hour or $45,760.

The unfortunate thing about minimum wage jobs is that you typically don't work 40 hours per week, in fact places like Wal-Mart try to keep employees under 30 hours per week to avoid the requirement of paying any healthcare benefits (at least 1/3 of Walmart employees are limited to 28 hours/week). This requires minimum wage employees to have two or three jobs which creates an extra layer of uncertainty for scheduling of work and getting from one location to another using public transportation. (as a side note WalMart is holding a food drive for their employees)

Why don't they move to another location that pays better for a job though? There are a couple reasons, first moving is expensive. Even if you're renting on both ends a cross country move costs several thousand dollars. You have to put money up front for a deposit, go without pay for some period of time as you transition between jobs, you need to move your stuff, and finally, many places won't hire you if you don't live there, but you can't rent unless you have a job. It's even worse if you own a house, and the worst case is a house that's underwater.

Now, under utility theory all that won't matter because it's a sunk cost. It's the cost of doing business, all that matters is the amount of money you'll gain on the other side. Now, if we were all perfectly rational "Econs" that would work, however we're Human and we don't think that way. According to Prospect Theory we are extremely loss adverse and would much rather have a small chance at a with with a worse downside than a guaranteed downside. For example, we'd rather have a 10% of winning $0, but 90% of losing $1,000 than automatically losing $900. From Utility theory or expected value they are the same, but from Prospect theory they are very different. This drives our behavior when it comes to relocating for jobs, and is especially true when you're unemployed. Because your baseline is set at your previous salary, you'd rather wait and risk making no money than taking a job that's more than 10% less than your previous salary. From an "Econs" perspective this is irrational as you should take any job that's offered to you, because it's more than $0. This theory can be applied to corporations as well. They are risk adverse in many cases. In one way this manifests is through the controllable costs of employees salaries. In many cases managers will let employees go or hire employees at a lower salary than deal with something that's a sunk cost.

Prospect Theory FourFold, Thinking, Fast and Slow Daniel Kahneman
Furthermore, corporations are under pressure do to short term requirements of the market to continually beat the previous quarters. This pushes activities in to the lower right quadrant, which also negatively impacts innovation and employee salary.

Based on the rationality here, should we increase minimum wage? Based on both Utility theory and Prospect theory companies would dramatically reduce the number of employees they have to avoid a sudden massive loss. Companies should move or relocate elsewhere, however, many studies have shown this simply isn't the case. I think that this can be explained by something else that's unexpected, unemployment benefits help the economy grow. According to a Moody's study several years ago and another talk a single dollar of unemployment benefits creates between $1.64 - $2.00 of economic activity. This is because it creates demand that otherwise would have been missing. It's likely that the higher salary of people that are making minimum wage generates additional demand for goods and services that otherwise would have been unable to be fulfilled.

I believe, this is why Switzerland is looking into providing a living wage of $2,800/month or $33,600/year. It's a way of being both moral (not allowing citizens to starve and live on the street), but also stimulate the economy to continually grow. The idea of lower taxes is to put more money into the economy so people will spend more money. However, there are a lot of people out there that don't benefit from this all that much because they already pay little to no taxes (40% before 2007). Providing tax relief doesn't help this portion of the economy. Providing a higher minimum wage would provide relief for them, but a tax on corporations, but would be recouped through additional demand. A living wage provides relief, but a tax on the broader public, but would be recouped through additional demand.

In the first case, people are working and providing a service to a corporation for the salary they earn. In the second case, there's a higher chance of free loading - however I see no reason why you can't put requirements on accepting the money such as community service, using the money for education and training, or volunteering somewhere.

Do these thoughts address every issues we've seen regarding minimum wage? No, but it's helps frame the discussion a little bit and hopefully addresses some of the concerns. I believe that much of debate comes from non-rational sources, such as the economic theory you fully support. Prospect theory, Behavioral Economics, and Evolutionary Economics are disdained or unknown by Libertarians and Conservatives which paint a very different picture on the economic policies we should enact than neo-classical economics. Which puts this to something closer of a religious battle than a logical rational debate, because these theories are incommensurate with each other in the same way that Newtonian Physics and Relativistic Physics were to each other. It's either one or the other not both.

Edit: I miscalculated on the salaries for minimum wages. I have corrected them.