Here's another example that I want to provide for you. This is this is a book that we published that I was part of a team that published when I worked at the CDC. This is the Incidence and Economic Burden of Injuries in the United States, and injuries include unintentional injuries and intentional injuries. And the intentional injuries is what I'm presenting for you here, which is the violence episode. So the background here is the data and the methods represent, the culmination of a three-year study. These, when this book was published in, I think 2006, yes, this was the most complete and thorough examination of national costs of injury since 1989. If you're interested in looking at more recent data than what I'm presenting here, the CDC has updated these figures on their website, and they're, they actually have an interactive tool that allows you to drill down, what are the economic costs of injuries and violence at the state level as well. So just to give you a summary of this study, this was, a summary was published in 2007 in AJPM. We looked at all injuries that occurred in the year 2000, we used a number of data sources to look at incidents. So we looked at the medically treated injuries, again, only injuries occurring in 2000. And remember, there are lots of injuries, particularly around violence, that occur in the year 2000, but never get reported as violence. So, those are not included here. They may end up showing up in the unintentional injuries section when they're really violent. So we do have a little bit of a reporting issue here on the violence. We also attempted to have mutually exclusive categories of injuries. We don't want to double count if someone uses an emergency department and they're have an inpatient admission and they end up dying. You don't want that person showing up in all three categories of illness. So we did, we took some great effort to try to make sure that we had mutually exclusive categories. And we looked at incidence by age, sex, mechanism, body region, nature and severity. So this is a cost of illness analysis. As we're talking about today, the medical costs are those for which claims exist. As we mentioned before, productivity losses, are exactly how I've defined them before, which is looking a lost time from work with some adjustments for household productivity. These are incidence based estimates which means that we are looking at lifetime costs. Although I will admit that the data on longitudinal costs associated with violence are pretty slim. So we really had a three year time horizon for looking up lifetime cost. And for those severe injuries severe, violent episodes that resulted in very severe injuries like traumatic brain injury, we were able to go to the literature and pull the lifetime cost of those severe injuries. We used, for those cost estimates that are unit costs, that we gathered in different time periods, we inflated everything to $2,000, and we used a 3% discount rate, to come up with net present value. The sources on the medical costs are, again, the typical sources that one uses to do national economic burden estimates, HCUP, Health Care Utilization Project, The Med. The market scan database, which is a wonderful database that includes claims from private employers. Has millions of covers of, covered lives, it's very helpful in doing these types of analyses the maps data, as I mentioned before. So we had 15 ICD diagnosis groupings used to estimate the cost by treatment location. We covered all injury diagnoses. again, we looked at body region, nature of injury or both. And we had to, do a little bit, be a little bit careful in some of the data sets that we used to make sure we, we're not pulling for, from some samples that were too small. How did we come up with our cost estimates for acute costs? We looked at inpatient facility costs using HCUP, we multiplied that times what is the unit cost for an inpatient admission, and we divided that by total costs. You can see how we came up with short to medium costs, which is up to 18 months. We have acute cost ratio from MarketScan from 1.03 to 1.39. this, basically is just taking data that we have from one data set and applying it to another so we can make adjustments for short term costs to long term costs. There's a very old database that we've, we're using which is called this detailed claim information set that had some longitudinal information on what are some costs beyond 18 months. And so we use that to make some adjustments to some of the other data sets that we had. All right, so, from this study we estimated that there were approximately 2.5 million violence-related injuries that occurred in the year 2000, 2.5 million. That number should not just go right over your head. That is an important, number because of the, of, of, because of how large it is. And if you look at the differences between fatal, and fatal assaults, versus non-fatal assaults, versus fatal self-inflicted wounds, versus total self-inflected wounds, you also see some important findings there. Particularly around the differences between men and women. So for fatal assaults, men experienced 12,000 fatal assaults, women only 4,000. So three times times the number of fatal assaults in men than in women. If you look at fatal self-inflicted, men are much more likely to a fatal self-inflicted wound, well, relative to women. and, but when it comes to total self-inflicted, females have about the same number as men. Why is that? It's because of the mechanism. Because men are using guns and women are using poison when it comes to self-inflicted wounds, and guns kill and poison makes you sick. And I think we might have some slides on that in a second. What are the total lifetime costs associated with those two, two million violent injuries? $70 billion for lifetime costs of those injuries that occurred in 2000. And again, because the incidence is so much higher for men than it is for women, the costs are running so much higher for men compared to women as well. Now, we can also use these types of data to help us, estimate the cost per event. So when you are thinking about doing interventions to prevent a fatal homicide or to prevent a suicide, one would want to know what is the cost associated with preventing that suicide. So these estimates here are very important for the field, for the, for research. So you can see the costs of a homicide is $1.3 million in lost productivity. That's in comparison to the cost of a suicide, which is one million. So why is there a difference in the value of statistical life between a homicide and a suicide? Well, it could have to do with the age that the suicide or the homicide occurred, It could have to do with whether the homicide is for a man or a woman. There's lots of things that go into that value of statistical life or the law future productivity stream as I mentioned before. And then you can see that the, the medical costs for homicides and suicides are relatively low because there's a fatality involved so there's no care. But the medical costs associated with non-fatal assaults and non-fatal self inflicted are fairly substantial. Now the gender differences in mechanism is what I was mentioning before, total costs were highest for males experiencing firearm assaults, which represented 52% of the total costs. Total costs were highest for females reporting struck by or against injuries. Again, as probably the victim of intimate partner violence. The rate of self-inflicted firearm injuries was six times higher for males than for females. And the rate of self-inflicted poisoning injury was 60% higher for females compared to males. Which, again, is driving that fatal versus nonfatal incidents rate. Now what are some of the challenges in assessing long term costs of violence? There are many. The first is that many of our data sets that I mentioned to you are older and potentially non-representative of the United States. There's some of the data that we had that were coming from only a few states, and we're applying it to national estimates, which may not be appropriate. We had to combine multiple data sources, which is always a great challenge. And there, and then, finally, and most importantly, there are very few longitudinal studies to determine long-term medical costs and productivity losses. In the field of child maltreatment in particular, we have good evidence like I showed you in the previous slide, that there are correlations, not causation, but correlations, between child maltreatment and then adult depression or adult obesity. But we've not done a good job in taking a cohort of abused kids and following them over their lifetime to and comparing them to a control group to really figure out what's going on in terms of causation and that then leads into how much the costs are. So the, the pros of the diagnosis-specific costs, which were these two examples that I just shared with you, is that it represents a lower bound on the actual costs of violence. Because again, we don't, since we're only doing diagnosis-specific, if there's a mental health claim, we won't know whether that mental health is tied to the violence or not. But, it is good for incidence based modeling. And the con is it may underestimate costs if co-morbid events are included, or not included and we know, at least from the literature, that are is some possibility of co-morbid events occurring. Now the final type of method that one might use in estimating cost of injury or cost of violence, is what's called the attributable fraction. And this comes from the field of epidemiology, where we include the indirect, where, in addition to including the medical costs and non-medical costs and productivity losses associated with the violent episode, we also try to get at, what are those indirect health expenditures associated with the other conditions or diseases that might be impacted, right? This is what we were trying to do in the previous methodology, this is what we said what we want to do in the previous methodology because we know it's an underestimate. So, in this case, you'd need to have some good epidemiologic data that shows, for this cohort of people who, who come into the doctors office, with the diagnosis of depression, we know that x percent of them were abused as children or something along those lines, and this gets us that attributable fraction. So I just wanted to share with you an example, from the literature. This comes from, from Australia, and you can see their website, there. They did an epidemiologic assessment of the health outcomes contributing to the disease burden of intimate partner violence in Victoria women. So they took some epidemiologic data, really at the very micro level to determine that 1% of STD, 1% of cancer or cervical cancer, et cetera, et cetera, are, are responsible because, are there because of the intimate partner violence that occurred. And so then what they did was they said, okay, the costs of STDs in Victoria is $500 million, so we're going to take 1% of that and say this is IPV related and we're going to put it in the cost of IPV pot, if you will. So the pros of doing this it directly deals with this idea that there are co-morbidities associated with violence. But the cons are pretty major and that is you have to have some pretty strong epidemiologic evidence from this little clinical trial right here to show that that the epi-evidence that you have can translate to the entire population. So it's moving from attributable fraction in one study to population attributable fraction which is a big epidemiologic hurdle to overcome. So what are, what are, are our gaps in knowledge about the economic impact of violence? Given these examples that I've showed you and the different ways of reporting. First, I would say that the first one is that we don't have a lot of information on attributable fraction and the economic impact. So while we, I did present that one. Example, there's, was a lot of criticism of that study when it was published and there just aren't a lot of good studies that are getting attributable fraction, particularly around depression,drug and alcohol use, and also obesity. Second I would that the impact on productivity needs some more research as well. I mentioned to you the research that we did with perpetrators of violence are those potential perpetrators of violence and their productivity losses and that was one of the first studies of its kind to look at productivity losses and the perpetrators. We have very little information on productivity losses for victims but it's probably important as a society for us to look at productivity losses for perpetrators as well, because they are having an impact. Third is the impact of intimate partner violence on children. So if you think about violence as a disease, which many people don't like to use that term but violence spills over into other areas. So it's not just the person who's being victimized, but it's the family members who witness that victimization. So there needs to be more work on what is the impact of intimate partner violence, domestic abuse or even, violence within the community on the children, and then how does that translate into long-term costs. And then finally, this incidence-based model to estimate long-term costs of child maltreatment is really needed. We need to be able to follow cohorts of children over time to see what are some of the economic impacts of experiencing abuse at such a, such a young age. There are bigger gaps in the field as well. Economic impact of violence that we've talked about today is really just the first step in assessing the economic impact. What we really need to do is conduct economic evaluations where we look at our interventions and we say, here's the potential cost savings by preventing these injuries, and here's how much the program costs. Looking at those two things together allows us to assess the efficiency of our scarce resources. It allows us to assess whether we have positive returns on investment. And that really brings us into the world of economic evaluation which is the next step after, cost of illness. So with that, I thank you for your attention, and if you have any questions I encourage you to submit them to the discussion forum. Thank you.