Monday, March 12, 2007
IS 21% CORRECT?
The health issues news service I use, care2.com, had an article posted today that claimed the UK has had a 21% increase in maternal childbirth deaths over the past year. It also claimed that the average salary for physicians in the UK is now higher than in the US, thanks to a new agreement between physicians and the NHS, whereas there is a hiring FREEZE for midwifes.
As the late David Falk would have said, interesting, if true. My problem is that I have found it unwise to take press reports(even those vetted by well-meaning folks like those at care2.com) at face value. After all, the much admired New York Times assured us a few years ago that Iraq possessed weapons of mass destruction.
Are these statistics accurate? Are there interesting issues here (along with the usual costly bout of Humean sympathy I am so prone to) for consideration under the topic of, say, "justice in the distribution of health care resources?" Will this be the occasion on which I loose those rose colored glasses through which I view all things "English?"
Bob Newsom ¶ 9:20 AM
Comments:
<< Home
Ah Bob.
Delightful questions as always and, at long last, something I actually think I am able to discuss without stumbling over my own feet along the way.
I recently wrote book reviews for two pieces that are germane. One was John Glad's “Future Human Evolution: Eugenics in the Twenty-First Century” and the other was Michael Mandelstam's "Betraying the NHS: Health abandoned".
Mandelstam's book is a quite detailed description of the failures within the NHS and of local health entities. It applies to most modern health care systems so is a good read. Glad's book is a sober reminder of the willingness of some people to advocate 'purification' of the race - either by benign neglect as Mandelstam describes, or overt acts of preferential genetic selection.
So, to your first question - is it possible that there has been an increase in maternal childbirth deaths in the NHS facilities? Yes, and for at least two perfectly good reasons. One is because we try to save more high risk pregnancies than we used to because we think we can technologically. Second, and more to the point you are probably raising, because there are increased problems in funding and providing services at local levels from communities to individual patient's bedsides, which might mean that services for high risk pregnancies are not available in all jurisdictions, or that needed services are limited/denied because of funding inadequacies.
Second, is it possible that NHS physicians are being paid more than US physicians. Certainly. Even tossing aside the Iraq induced inflation, the unfunded long term costs of caring for Iraq vets, and the drug induced spending frenzy of the current inhabitant of the WH which has left us trillions of dollars more long term debt than 10 years ago, with its attendant shortfall on value of physician's salaries, the structure of physician practice in the US has changed dramatically the last couple of decades though most of us would probably prefer to think that it has not. Many American physicians are wage employees rather than solo practitioners managing their own operations. New grads may be working in managed care settings, directly for integrated health care delivery systems, or they may be seeking partnerships in ongoing practices much the way lawyers seek partnerships in law firms, accepting reduced income and high work loads their first few years as they build up 'equity' and hopefully become partners some day. Many health care providers in the US see physicians in much the same way that they view billing clerks and receptionists. Hire and fire them as they see how well they fit in the service delivery model embraced by the managers above them - generally seen as "X generates a lot more revenue than costs". Unlike physician's salaries in the 1950s and 1960s, when med school graduates came from upper middle class families and often became partners, American physician's salaries have severely eroded as a result of prospective payment systems, capitation contracts, and retrospective audits all of which tend to reduce the amounts that providers of health care services actually receive. Factor in inflation and the value of the dollar in international markets, and the statement makes sense.
As though those issues were not enough, there are other finance problems that all modern health care systems and care providers face that give rise to a great many opportunities for philosophical as well as financial debate. The most pervasive method of health care finance in use today involves implicit or explicit transfers of health insurance risks from large entities (we used to call them health insurers but there are few of them remaining) to small entities. The classic mechanism is a capitation contract. In return for a guaranteed monthly per capita payment, health care providers promise to provide the unknown future service needs of a defined cohort of patients. Each month the provider gets a payment from the 'insurer', and in return, delivers health care services.
Were it not for the fact that these arrangements violate the fallacy of distribution and extant insurance practices, not to mention the core insurance risk management explanations derived from probability theory, statistical sampling theory, mathematical risk theory, basic economics, and utility theory, there might not be any problem at all.
Unfortunately, these mechanisms - you might think of them as capitation agreements, Prospective Payment Systems, Diagnosis Related Groups, utilization review, budgeted services such as nursing, or retrospective audits if you like, do violate the rules of insurance. On a parallel course, you might think in terms of Individual Health Accounts, consumer directed health care, or local control initiatives - the things being pushed so hard by the WH...
Why are these things bad? The answer is pretty simple. All these mechanisms have the common trait of transferring both clinical risks - the management of patients health care needs AND financial risks - the management of the health insurance liabilities for the patient cohorts involved - to health care providers. The health care providers assume dual roles as care providers and mini-insurers. One duty of insurers is claims handling - denying claims you are not obliged to pay and honoring those you are obliged to pay - though the obligation is often fuzzy and hence, it makes sense to deny first and ask whether it is justified later - think Katrina and Homeowners insurance...
But large insurers manage risk far better than small insurers because of the Central Limit Theorem and the Law of Large Numbers. In essence, an insurer with a million policyholders will have far more stable operating results, year to year, than an insurer with 100,000 policyholders because they take a larger sample of the population of all possible policyholders. Pull down an elementary statistics book from the top shelf to review, if necessary, how the standard error of the estimate of the mean loss ratio of a large insurer would compare with the standard error of the estimate of the mean loss ratio for the smaller insurer (Hint: square root of the ratio of sample sizes).
So, where the large insurer (1,000,000 policyholders) may be able to accurately estimate the range of losses it faces with a probability of 0.95 - let's say it runs between $0.60 and $0.90 per dollar of premium, the smaller insurer, insuring the identical risks, although fewer policyholders (100,000) will have a spread that is about 3 times wider merely because of the smaller sample size involved. From a practical standpoint, the smaller insurer (Now let's think of capitated health care provider) must target their average service delivery costs at a lower level than the large insurer because of the possibility that they will have a larger loss due to randomness than they can manage. In this example imagine that the providers loss ratio varies between $0.30 and $1.20 per dollar of premium collected by the large insurer depending on whether the flu hits their community. No flu - great year. Bad flu - higher costs than manageable - again, think Katrina and the care of nursing
home and hospital patients in the critical hours after Katrina hit. Of course, most providers will have results that are fairly close to the mean for the population at years end - but they cannot be sure of that at the start of the year or even during the year in which they are obliged to deliver services. If they cut back they may actually have a loss ratio of $0.20 in a very good year, but if they fail to cut back they may bankrupt themselves in a really bad year with losses that exceed the revenues from their contracts. There are tons of other problems as well - but that is too much to try to cover.
The bottom line? Health care providers that assume health insurance risks are inefficient insurers. To compensate for their inefficiency as insurers they have to reduce service quality (measured by cost) and quantity lest they run out of funds before the end of the contract period.
Some of the interesting philosophical/ethical issues: The fallacious reasoning that underlies common health care finance debates; The issues of whether health care providers ought to be assuming dual roles as caregivers and insurers; The issues regarding how providers make tradeoff decisions between clients - will I treat Ms. Brown's breast cancer aggressively or will I vaccinate 10,000 school children, especially if one alternative results in higher net profits to me but both are equally easy to sell to the public?
Most of our current discussion about health care finance assumes that individuals, groups of individuals, and communities will be able to manage health care costs for themselves with an amount that is, on average, adequate across the larger population. This is, of course, patently false.
Probably the simplest explanation I have prepared to date is my article in Nurse Leader last year. Pretty palatable unless you are a risk accepting provider, a politician, or an apologist for the managed care and insurance industries... A somewhat more detailed explanation is in my 2004 dissertation. Just look at the explanations of Professional Caregiver Insurance Risk, not the rest of the gobbledygook. The effects? Necessary reductions in care for the elderly and mentally ill - as described by Mandelstam, because it makes good sense to knock off services for small numbers of high cost clients first - greatest good for the greatest number and all that. As well, there are bound to be increased rates of poor outcomes, as care is delayed, deferred, or denied by health care providers that are stumbling over their financial risks and clinical obligations, and increased financial vulnerability for providers...
Cox,T. (2006). Professional caregiver insurance risk: A brief primer
for nurse executives and decision-makers. Nurse Leader, 4(2):
48-51.
Delightful questions as always and, at long last, something I actually think I am able to discuss without stumbling over my own feet along the way.
I recently wrote book reviews for two pieces that are germane. One was John Glad's “Future Human Evolution: Eugenics in the Twenty-First Century” and the other was Michael Mandelstam's "Betraying the NHS: Health abandoned".
Mandelstam's book is a quite detailed description of the failures within the NHS and of local health entities. It applies to most modern health care systems so is a good read. Glad's book is a sober reminder of the willingness of some people to advocate 'purification' of the race - either by benign neglect as Mandelstam describes, or overt acts of preferential genetic selection.
So, to your first question - is it possible that there has been an increase in maternal childbirth deaths in the NHS facilities? Yes, and for at least two perfectly good reasons. One is because we try to save more high risk pregnancies than we used to because we think we can technologically. Second, and more to the point you are probably raising, because there are increased problems in funding and providing services at local levels from communities to individual patient's bedsides, which might mean that services for high risk pregnancies are not available in all jurisdictions, or that needed services are limited/denied because of funding inadequacies.
Second, is it possible that NHS physicians are being paid more than US physicians. Certainly. Even tossing aside the Iraq induced inflation, the unfunded long term costs of caring for Iraq vets, and the drug induced spending frenzy of the current inhabitant of the WH which has left us trillions of dollars more long term debt than 10 years ago, with its attendant shortfall on value of physician's salaries, the structure of physician practice in the US has changed dramatically the last couple of decades though most of us would probably prefer to think that it has not. Many American physicians are wage employees rather than solo practitioners managing their own operations. New grads may be working in managed care settings, directly for integrated health care delivery systems, or they may be seeking partnerships in ongoing practices much the way lawyers seek partnerships in law firms, accepting reduced income and high work loads their first few years as they build up 'equity' and hopefully become partners some day. Many health care providers in the US see physicians in much the same way that they view billing clerks and receptionists. Hire and fire them as they see how well they fit in the service delivery model embraced by the managers above them - generally seen as "X generates a lot more revenue than costs". Unlike physician's salaries in the 1950s and 1960s, when med school graduates came from upper middle class families and often became partners, American physician's salaries have severely eroded as a result of prospective payment systems, capitation contracts, and retrospective audits all of which tend to reduce the amounts that providers of health care services actually receive. Factor in inflation and the value of the dollar in international markets, and the statement makes sense.
As though those issues were not enough, there are other finance problems that all modern health care systems and care providers face that give rise to a great many opportunities for philosophical as well as financial debate. The most pervasive method of health care finance in use today involves implicit or explicit transfers of health insurance risks from large entities (we used to call them health insurers but there are few of them remaining) to small entities. The classic mechanism is a capitation contract. In return for a guaranteed monthly per capita payment, health care providers promise to provide the unknown future service needs of a defined cohort of patients. Each month the provider gets a payment from the 'insurer', and in return, delivers health care services.
Were it not for the fact that these arrangements violate the fallacy of distribution and extant insurance practices, not to mention the core insurance risk management explanations derived from probability theory, statistical sampling theory, mathematical risk theory, basic economics, and utility theory, there might not be any problem at all.
Unfortunately, these mechanisms - you might think of them as capitation agreements, Prospective Payment Systems, Diagnosis Related Groups, utilization review, budgeted services such as nursing, or retrospective audits if you like, do violate the rules of insurance. On a parallel course, you might think in terms of Individual Health Accounts, consumer directed health care, or local control initiatives - the things being pushed so hard by the WH...
Why are these things bad? The answer is pretty simple. All these mechanisms have the common trait of transferring both clinical risks - the management of patients health care needs AND financial risks - the management of the health insurance liabilities for the patient cohorts involved - to health care providers. The health care providers assume dual roles as care providers and mini-insurers. One duty of insurers is claims handling - denying claims you are not obliged to pay and honoring those you are obliged to pay - though the obligation is often fuzzy and hence, it makes sense to deny first and ask whether it is justified later - think Katrina and Homeowners insurance...
But large insurers manage risk far better than small insurers because of the Central Limit Theorem and the Law of Large Numbers. In essence, an insurer with a million policyholders will have far more stable operating results, year to year, than an insurer with 100,000 policyholders because they take a larger sample of the population of all possible policyholders. Pull down an elementary statistics book from the top shelf to review, if necessary, how the standard error of the estimate of the mean loss ratio of a large insurer would compare with the standard error of the estimate of the mean loss ratio for the smaller insurer (Hint: square root of the ratio of sample sizes).
So, where the large insurer (1,000,000 policyholders) may be able to accurately estimate the range of losses it faces with a probability of 0.95 - let's say it runs between $0.60 and $0.90 per dollar of premium, the smaller insurer, insuring the identical risks, although fewer policyholders (100,000) will have a spread that is about 3 times wider merely because of the smaller sample size involved. From a practical standpoint, the smaller insurer (Now let's think of capitated health care provider) must target their average service delivery costs at a lower level than the large insurer because of the possibility that they will have a larger loss due to randomness than they can manage. In this example imagine that the providers loss ratio varies between $0.30 and $1.20 per dollar of premium collected by the large insurer depending on whether the flu hits their community. No flu - great year. Bad flu - higher costs than manageable - again, think Katrina and the care of nursing
home and hospital patients in the critical hours after Katrina hit. Of course, most providers will have results that are fairly close to the mean for the population at years end - but they cannot be sure of that at the start of the year or even during the year in which they are obliged to deliver services. If they cut back they may actually have a loss ratio of $0.20 in a very good year, but if they fail to cut back they may bankrupt themselves in a really bad year with losses that exceed the revenues from their contracts. There are tons of other problems as well - but that is too much to try to cover.
The bottom line? Health care providers that assume health insurance risks are inefficient insurers. To compensate for their inefficiency as insurers they have to reduce service quality (measured by cost) and quantity lest they run out of funds before the end of the contract period.
Some of the interesting philosophical/ethical issues: The fallacious reasoning that underlies common health care finance debates; The issues of whether health care providers ought to be assuming dual roles as caregivers and insurers; The issues regarding how providers make tradeoff decisions between clients - will I treat Ms. Brown's breast cancer aggressively or will I vaccinate 10,000 school children, especially if one alternative results in higher net profits to me but both are equally easy to sell to the public?
Most of our current discussion about health care finance assumes that individuals, groups of individuals, and communities will be able to manage health care costs for themselves with an amount that is, on average, adequate across the larger population. This is, of course, patently false.
Probably the simplest explanation I have prepared to date is my article in Nurse Leader last year. Pretty palatable unless you are a risk accepting provider, a politician, or an apologist for the managed care and insurance industries... A somewhat more detailed explanation is in my 2004 dissertation. Just look at the explanations of Professional Caregiver Insurance Risk, not the rest of the gobbledygook. The effects? Necessary reductions in care for the elderly and mentally ill - as described by Mandelstam, because it makes good sense to knock off services for small numbers of high cost clients first - greatest good for the greatest number and all that. As well, there are bound to be increased rates of poor outcomes, as care is delayed, deferred, or denied by health care providers that are stumbling over their financial risks and clinical obligations, and increased financial vulnerability for providers...
Cox,T. (2006). Professional caregiver insurance risk: A brief primer
for nurse executives and decision-makers. Nurse Leader, 4(2):
48-51.
# posted by 
: 9:37 AM
: 9:37 AM
A very good explanantion and feedback Bear, though it is dissapointing in terms of change that numbers are seen to be the defining characterisitic regarding health. You mention some particular vulnerable groups and yet if given some intenstive resources ( ie upstream work) then further issues down stream may be avoided. Whilst risks are inevitable, reframing health into just risk scenarios is not necessrily going to alter perceptions of what health is or can be.
Given that today is international womens day - perhaps a word from a bill poster - by the age of 55 more women than men have high cholesterol. Now to take Bobs point, interpreting facts from various sources such as the media may require addtional scrutiny. However, other sources too ( such as news/stats sites etc) as you say require critical analysis and debate.
The dual role bit does beg questioning yet in some cases that is it - maybe more rethink on what it is insurance is or ought to be doing - ethical business and the like.
Given that today is international womens day - perhaps a word from a bill poster - by the age of 55 more women than men have high cholesterol. Now to take Bobs point, interpreting facts from various sources such as the media may require addtional scrutiny. However, other sources too ( such as news/stats sites etc) as you say require critical analysis and debate.
The dual role bit does beg questioning yet in some cases that is it - maybe more rethink on what it is insurance is or ought to be doing - ethical business and the like.
# posted by : 11:35 AM
: 11:35 AM
Hi Sarah,
Points well taken.
PCIR, as I like to think of it, is a uniquely 'nursing' analysis of health care finance because of its explicit concern with the quantity and quality of service available to clients downstream. In my most grandiose days I imagine Flo would have done the same analysis, albeit a lot sooner and better than I have to date. So, keep in mind that my thinking is not exclusively financial, even though it is necessarily financial, because I am a nurse and because the situation in terms of health care policy and finance debates demands that nurses engage in financial analysis if we are to honor our obligations to our clients rather than be caught blind-sided over and over, imagining each new financial deception imposed on us as fundamentally different from the last.
The real key, which I clearly did not communicate well enough, is that the impact of insurance risk assumption by health care providers necessarily results in service inadequacies down stream. Let me try to flesh that out a bit better.
Suppose we had a perfectly efficient health care system and an efficient insurance market place. Making this assumption helps us avoid the flaws of thinking that if we just 'adjust' something it will improve the situation. We assume that all possible adjustments have already been made and the benefits taken into account.
In our efficient system everyone gets treated at the best level for their needs - so we can't duck problems at the local level on the basis that we could send patients somewhere else. Focusing on mental health services, we would know, quite accurately, the costs of providing a specific set of mental health care services to the average citizen - some, of course, use no mental health services, while others are big users - frequent medication vacations, hospitalizations, re-entry issues. At the national level we know what the costs are per citizen. So we collect funds adequate to meet those costs and these funds are transferred, in the form of capitation agreement payments, to a local mental health board for local services. Obviously we want the local mental health board to be efficient, and one measure of efficiency is actual cost compared to allocated cost. Many other forms of efficiency come to mind - but remember that we have already achieved all the possible benefits of service efficiency, and what we are left with is how much money the local board gets and how much it spends.
Using my earlier example, let's assume that the cost of mental health services nationally is $0.75 per dollar of revenue for mental health services coverage. At the national level we know that the amount actually spent, varies between $0.60 and $0.90 (standard error = $0.15/1.96) per dollar of mental health revenues each year, though it is usually close to $0.75. Now, because the local community demographics are not precisely the same as the national demographics, we have either more of less use of mental health services locally than nationally. As well, if there is a local plant closure (More nursing perspective on environment) we can anticipate higher levels of depression and anxiety, and higher costs of treatment at some local facilities and lower costs where the plant is re-opened in a lower cost of labor area. The difference between the local community and national figures is one sort of risk involved - process mean risk. If the local area has more vulnerable mentally ill our average cost will be higher than average nationally. If our local community is a lot healthier than average, we will have lower costs than nationally.
Thus far we have not, however, dealt with the real problem Professional Caregiver Insurance Risk addresses - random variation due to the unpredictability of exactly who will use mental health services during the year. If we were selecting our local population at random from 10% of the national population, our standard error of the estimate of the mean loss ratio would be 3.162 times the standard error of the estimate of the mean loss ratio for the national population. This means that if we think in terms of a normal distribution curves the smaller insurer will have much fatter tails than the larger insurer. The probability for any actual loss ration higher than $0.75 will be higher for the small insurer than for the large insurer.
Our 'expected' costs for mental health services would be precisely the same as the expected cost for the nation, again assuming random selection. However, the smaller entity will see variation around the expected value that runs between $0.28 - $1.12 ($0.75 +/-3.162*$0.15) solely due to the randomness in the cohort selected, due to the insurance risk transfer involved, and the size differential between the large insurer and the small insurer. I am using big numbers, in reality the standard errors would be measured in fractions of a cent but the relative sizes are what really counts. Again here, we have achieved all the efficiencies possible so there is no redundancy in the system, since that would be inefficient. Where costs in excess of $0.90 at the national level occur, by design, in only about 2.5% of years for the national entity, the losses in excess of $0.90 of larger insurer premium occur at the local level far more often (slightly less than 50% of the time) because of the larger standard error and the flatter normal curve involved which results in far heavier tails.
If the local entity wants to make sure it does not exceed its allotment - which by agreement cannot exceed the expected value of $0.75 taken in nationally, and given its much larger standard error, it would have to target service delivery levels far below $0.75 if it wants to assure solvency at the end of the year. If it wants to limit its losses to be no higher than its revenues from the large insurer except 2.5% of the time, it would have to adjust its expected costs down low enough to compensate for the larger standard error. How far? Well we already have our offsets to the mean - they are $0.47 for the inefficient insurer. So if it targets costs at $0.28 it will not exceed its revenues of $0.75 except 2.5% of the time. Only problem is - this targeted service delivery level is far below the service levels it delivered before it went into the insurance business.
Since we think we need to make drastic cuts - and we do because we are inefficient insurers - we will, at the local level, proceed as detailed in Mandelstam's book, we might as well look for those rare, high cost services, figuring that we will harm the fewest people and benefit the most. In a good situation we find a tiny fraction of patients who generate very large costs, and who are politically powerless, such as the aged infirm and the acutely mentally ill, cut their services completely, and apologize for our failure to recognize that their needs were unmet, after they are all thrown out of hospitals, driven to skid row in cabs, or otherwise 'disappeared' into the general population.
The rationale here is exactly that used by insurers in settling very large claims. If you tap the car in front of you while driving and do $1,000 of damage, your insurance company will pay lickety split. But if you cruise into that car at 60 mph and permanently disable 4 people who will generate health care costs of $500,000 your insurer will spend 4 years investigating the accident before it agrees to pay. Of course, in those four years two of the victims will die, and the other two will lose their homes and become wards of the state - but that is how insurers operate. Why? Because a very small number of people generate high claim costs and the general public is easily persuaded by arguments that the victims do not deserve such large awards and that self-serving greedy lawyers are the real culprits. When your physician/nurse and insurer are the same person it is like it is now... bad.
Given the fact that we have succeeded all too well in the past at preserving human life, regardless of the costs involved, blaming poor service delivery on mistakes at the local level is a lot more politically palatable than admitting that we are rationing health care services...
Points well taken.
PCIR, as I like to think of it, is a uniquely 'nursing' analysis of health care finance because of its explicit concern with the quantity and quality of service available to clients downstream. In my most grandiose days I imagine Flo would have done the same analysis, albeit a lot sooner and better than I have to date. So, keep in mind that my thinking is not exclusively financial, even though it is necessarily financial, because I am a nurse and because the situation in terms of health care policy and finance debates demands that nurses engage in financial analysis if we are to honor our obligations to our clients rather than be caught blind-sided over and over, imagining each new financial deception imposed on us as fundamentally different from the last.
The real key, which I clearly did not communicate well enough, is that the impact of insurance risk assumption by health care providers necessarily results in service inadequacies down stream. Let me try to flesh that out a bit better.
Suppose we had a perfectly efficient health care system and an efficient insurance market place. Making this assumption helps us avoid the flaws of thinking that if we just 'adjust' something it will improve the situation. We assume that all possible adjustments have already been made and the benefits taken into account.
In our efficient system everyone gets treated at the best level for their needs - so we can't duck problems at the local level on the basis that we could send patients somewhere else. Focusing on mental health services, we would know, quite accurately, the costs of providing a specific set of mental health care services to the average citizen - some, of course, use no mental health services, while others are big users - frequent medication vacations, hospitalizations, re-entry issues. At the national level we know what the costs are per citizen. So we collect funds adequate to meet those costs and these funds are transferred, in the form of capitation agreement payments, to a local mental health board for local services. Obviously we want the local mental health board to be efficient, and one measure of efficiency is actual cost compared to allocated cost. Many other forms of efficiency come to mind - but remember that we have already achieved all the possible benefits of service efficiency, and what we are left with is how much money the local board gets and how much it spends.
Using my earlier example, let's assume that the cost of mental health services nationally is $0.75 per dollar of revenue for mental health services coverage. At the national level we know that the amount actually spent, varies between $0.60 and $0.90 (standard error = $0.15/1.96) per dollar of mental health revenues each year, though it is usually close to $0.75. Now, because the local community demographics are not precisely the same as the national demographics, we have either more of less use of mental health services locally than nationally. As well, if there is a local plant closure (More nursing perspective on environment) we can anticipate higher levels of depression and anxiety, and higher costs of treatment at some local facilities and lower costs where the plant is re-opened in a lower cost of labor area. The difference between the local community and national figures is one sort of risk involved - process mean risk. If the local area has more vulnerable mentally ill our average cost will be higher than average nationally. If our local community is a lot healthier than average, we will have lower costs than nationally.
Thus far we have not, however, dealt with the real problem Professional Caregiver Insurance Risk addresses - random variation due to the unpredictability of exactly who will use mental health services during the year. If we were selecting our local population at random from 10% of the national population, our standard error of the estimate of the mean loss ratio would be 3.162 times the standard error of the estimate of the mean loss ratio for the national population. This means that if we think in terms of a normal distribution curves the smaller insurer will have much fatter tails than the larger insurer. The probability for any actual loss ration higher than $0.75 will be higher for the small insurer than for the large insurer.
Our 'expected' costs for mental health services would be precisely the same as the expected cost for the nation, again assuming random selection. However, the smaller entity will see variation around the expected value that runs between $0.28 - $1.12 ($0.75 +/-3.162*$0.15) solely due to the randomness in the cohort selected, due to the insurance risk transfer involved, and the size differential between the large insurer and the small insurer. I am using big numbers, in reality the standard errors would be measured in fractions of a cent but the relative sizes are what really counts. Again here, we have achieved all the efficiencies possible so there is no redundancy in the system, since that would be inefficient. Where costs in excess of $0.90 at the national level occur, by design, in only about 2.5% of years for the national entity, the losses in excess of $0.90 of larger insurer premium occur at the local level far more often (slightly less than 50% of the time) because of the larger standard error and the flatter normal curve involved which results in far heavier tails.
If the local entity wants to make sure it does not exceed its allotment - which by agreement cannot exceed the expected value of $0.75 taken in nationally, and given its much larger standard error, it would have to target service delivery levels far below $0.75 if it wants to assure solvency at the end of the year. If it wants to limit its losses to be no higher than its revenues from the large insurer except 2.5% of the time, it would have to adjust its expected costs down low enough to compensate for the larger standard error. How far? Well we already have our offsets to the mean - they are $0.47 for the inefficient insurer. So if it targets costs at $0.28 it will not exceed its revenues of $0.75 except 2.5% of the time. Only problem is - this targeted service delivery level is far below the service levels it delivered before it went into the insurance business.
Since we think we need to make drastic cuts - and we do because we are inefficient insurers - we will, at the local level, proceed as detailed in Mandelstam's book, we might as well look for those rare, high cost services, figuring that we will harm the fewest people and benefit the most. In a good situation we find a tiny fraction of patients who generate very large costs, and who are politically powerless, such as the aged infirm and the acutely mentally ill, cut their services completely, and apologize for our failure to recognize that their needs were unmet, after they are all thrown out of hospitals, driven to skid row in cabs, or otherwise 'disappeared' into the general population.
The rationale here is exactly that used by insurers in settling very large claims. If you tap the car in front of you while driving and do $1,000 of damage, your insurance company will pay lickety split. But if you cruise into that car at 60 mph and permanently disable 4 people who will generate health care costs of $500,000 your insurer will spend 4 years investigating the accident before it agrees to pay. Of course, in those four years two of the victims will die, and the other two will lose their homes and become wards of the state - but that is how insurers operate. Why? Because a very small number of people generate high claim costs and the general public is easily persuaded by arguments that the victims do not deserve such large awards and that self-serving greedy lawyers are the real culprits. When your physician/nurse and insurer are the same person it is like it is now... bad.
Given the fact that we have succeeded all too well in the past at preserving human life, regardless of the costs involved, blaming poor service delivery on mistakes at the local level is a lot more politically palatable than admitting that we are rationing health care services...
# posted by : 11:36 AM
: 11:36 AM
Tom:
You have outdone yourself. Someone a few years ago said of a passage from Aristotle that a teaspoon of it weighed more than a universe. I think you have provided us with another example.
Tons of social and political philosophy issues raised by your missive, I am sure, but I need a bit of time to digest it.
None of the concepts are foreign to me, nor are they unique to healthcare. All of these risk-sharing devices are well understood in a general way; all are in use, in one form or another, in, for example, the oil industry (which I am quite familiar with).
Come to think of it, the parallels are fascinating, beginning with the most obvious; energy policy gets "formulated" by politicians who haven't the foggiest notion of how the real world of energy exploration, production and delivery actually works. If the Carter administration, for instance, had understood ANY of the details, it would have understood that a "windfall profits" tax was simply going to end up being an excise tax on petroleum products that consumers, not producers, would end up paying. I was in Conoco's legal department back then, and we just read the damn thing and snickered. We got the windfall, YOU paid the tax.
Wasn't I just a prince of person back then? Then along came all you wonderful nurses, and have changed me in to the absolute saint I am today. Just HOW that was accomplished is, of course, another subject entirely. My Dundee presentation, pehaps.
Bob
Post a Comment
You have outdone yourself. Someone a few years ago said of a passage from Aristotle that a teaspoon of it weighed more than a universe. I think you have provided us with another example.
Tons of social and political philosophy issues raised by your missive, I am sure, but I need a bit of time to digest it.
None of the concepts are foreign to me, nor are they unique to healthcare. All of these risk-sharing devices are well understood in a general way; all are in use, in one form or another, in, for example, the oil industry (which I am quite familiar with).
Come to think of it, the parallels are fascinating, beginning with the most obvious; energy policy gets "formulated" by politicians who haven't the foggiest notion of how the real world of energy exploration, production and delivery actually works. If the Carter administration, for instance, had understood ANY of the details, it would have understood that a "windfall profits" tax was simply going to end up being an excise tax on petroleum products that consumers, not producers, would end up paying. I was in Conoco's legal department back then, and we just read the damn thing and snickered. We got the windfall, YOU paid the tax.
Wasn't I just a prince of person back then? Then along came all you wonderful nurses, and have changed me in to the absolute saint I am today. Just HOW that was accomplished is, of course, another subject entirely. My Dundee presentation, pehaps.
Bob
Subscribe to Post Comments [Atom]
<< Home
