Welcome, this is David Bishai, and in this section, we're going to be going over a more advanced model of volume and quality. And I wanted to start by reminding ourselves where we were at the end of the last section. We saw that volume responses to quality really dampened our ability to improve quality. When we set that quality constant up high, as volume constantly kept on going higher, we could not get ahead. So, the system improvements that bring quality at a low volume will fail when the volume increases. And that is a property of the system. As long as we believe that there is negative feedback between volume and quality, that will be a problem. Now of course, this is a simulation. Is it real? Is this a part of healthcare? Well, you can ask stakeholders. After you ask stakeholders what to do, you might then proceed to get more modeling done and more data collection and get policy makers to focus on the right thing. So for a minute, I'm going to stop being a systems dynamic modeler and show you my other life, which is to be a health care provider. I work once a week in an emergency room, and let me show you my dashboard. My manager in the emergency room sends me this email once a month to tell me how I'm doing in my work as an emergency doctor. This slide shows a call out button showing me, you can see that little red bar in the middle that says Y-O-U, well that actually means me. And that's my rating in comparison to all of my colleagues in the emergency department. And then those little gauges on the right are telling me how I was doing in this particular month. The RVU per hour indicator is how many relative value units I produce every hour of work. The RVU is what my manager cares a lot about because the RVU's are worth money. We go to the insurance companies and tell them how many RVU's we gave their patients, and that turns into the fee. So the manager's telling me I'm in the middle of the indicator, in the yellow zone of my RVU's per hour. I'm doing green zone on RVU's per patient, I'm scoring about 3.6 on RVU's per patient. I'm seeing 1.6 patients per hour. The PatSat indicator is Patient Satisfaction and I'm in the green zone on Patient Satisfaction at 4.49 and my percent of five ratings, five is the top scale, I'm getting 70% fives from my patients. And I guess the manager didn't put in the data for my length of stay because the length of stay indicator is parked at the worst possible indicator. And I know that's not me so ignore that length of stay indicator. But I'm telling you that this is how my manager tries to improve my quality and my manager is not paying any attention to the type of model we did in section one. Whereas the volume, I have to deal with increases, they don't count that as something that should affect my ability to perform quality. They don't adjust my patient satisfaction score for quality, and they simply try to push the indicators that they care about. Obviously, they care about RVU's per hour and that's why they're incentivizing me. So my manager sends me this indicator and they give me monthly reports of where I am. The manager is trying to appeal to my intrinsic motivation to be a good doctor. And in the diagram, they're showing me what they seem to care about, which is RVU's per hour, RVU's per patient, and patients per hour in this main bar graph indicator. Now I know, as a doctor, that patients per hour makes it very hard to have high quality. It's hard to spend enough time with a patient, to sit down and listen to everything they have to say. And spent time in the room explaining everything that we're doing is the essence of good physician services. When it's crowded, I have to keep moving to keep the waiting room from getting overcrowded. And the manager isn't using this type of information as they try to manage the system. So let's go to a more advanced model that my manager might use in order to try to get both quality and volume to work together. And this advanced model is available in Vincent for us. It's a model where they might used performance based financing in order to drive the motivation of the doctor's to produce better quality. So, at the bottom of this advanced model is this same set of loops we saw before, between technical quality and volume. The exact same as before at the bottom. But now, we have some feedback going on. We have the quality game being influenced by extrinsic motivation, meaning the motivation that comes from outside the doctor. And that is a state. It has an inflow motivation gain and an outflow motivation loss. And the motivation gain is going to be driven by the green part of the model, where there's performance based financing. There's this stock of performance based financing money at the top. Money flows into that pool of performance based financing money, and money flows out as disbursements. And the disbursements essentially go to the doctor in proportion to their quality or volume, and that reward of a bonus is supposed to give the doctor motivation to improve their quality. Now that motivation has a rate of loss and the rate of loss you can see if you look in that red circle, the motivation loss is driven by volume. So again, we have that negative feedback. High volume makes the doctor lose their motivation to try to produce high quality, and it increases the rate of quality loss, as we saw in the prior model. So we still have negative feedback when there's high volume. And, the question is, can you keep up with it with the performance based financing money, can we drive the system towards quality if we spend enough money on waiving bonuses at the doctors? So, that's what's in the model and you can actually open this model in Vincent and I encourage you to do that. And when you do, you can click on the settings of the model. You can change that rate of external funding coming into the performance based financing. And this is what will happen if you. In the left of the slide is a run of the model where we've put the external funding for bonuses and performance-based financing at a low rate of 0.5, and in that setting of the model, we get that same problem. Quality goes up at the beginning, volume responds to it and pushes down the quality, and we end the model with low quality at 0.25. So how about we triple the amount of external funding. We run it not at 0.5 but at 1.5. This maxing out our performance-based financing strategy. Well, when we do that, sure, you end with an equilibrium with just a little bit more quality, instead of ending up at about 0.3, now this ends up at about .35 or .4, not very impressive. And it costs a lot of money. We've spent a lot on trying to motivate them with offers of bonuses and we still haven't gotten what we want out of the system. So, what else to do? Well, focus on the negative feedback. We have this essential problem of the drag on the quality from that negative feedback from volume. So, I've circled in red that essential insight we have on the system. That it's the negative feedback that has to be overcome right there. And think about what could be done at that rate of motivation loss. And we can ask ourselves, would it matter if we could do policies on the rate of motivation loss, would they be effective? Since this is a simulation, we can answer that question using the computer simulation. And this is an area of greater impact, as we can see in the simulation. At a very low rate of motivation loss when we set that number at 0.3, we can end up with an equilibrium quality that's above 0.5. It's looking like 0.55 at the end of the model on the left. That blue curve of quality ends up quite high. But if we set up the rate of motivation loss at a high level, at 1.5, we end up with low quality at the end of the model on the right, of .2. So in the simulation, we can see that that rate of motivation loss could have a lot more impact on the system, it would make us work harder at thinking through policies that could be used to control the rate at which volume destroys the motivation of the doctor to produce high quality. Now we'll consider what polices could make the providers immune to higher volume. Now some of this has already been discovered by healthcare managers. If you go to an emergency room or any clinic, you will see ancillary providers that are helping the doctors and providers respond to high volume. They'll have the ancillary workers take the vital signs, take the medication history, put the patients in the rooms. This is all helping that volume come through without taking down the motivation of the providers. Recently, we're seeing the introduction of scribes to help with the charting. So instead of the doctor having to spend hours on every shift writing out the chart, the scribe shadows the doctor, listens to what the doctor is saying, listens to the physical exam and also creates a reduction in the burden that might reduce the quality. So, scribes are being tried to immunize providers. And obviously electronic medical records and the e-health and m-health revolution is, again, an experimental approach, which we're still working out, on how one can have clinics where there's high volume. But we still can preserve the quality response, then take away that reduction in the motivation of the doctors. So in our advanced model, we've just started to scratch the surface of all of the things you could do to gain insight into policies that could be used to control our key problem of getting high quality in our clinics and responding to high volume. In the next section, we're going to shift gears and take an even more macro picture at the entire health system.
