Archive for Disease Management
by JParkinson
April 21, 2008 at 11:41 am · Filed under Disease Management, Business of Health
If asthma takes $200 to prevent and $10,000 to treat, why would any self-respecting doctor or hospital want to prevent such a lucrative opportunity?
Why would any hospital group want to use Bayesian statistics to intelligently use radiology studies when intelligent ordering decreases revenue by 30%?
Why would we want to cut off the hand that feeds us?
Healthcare is a business. Profound, I know. Until we implement policies that pay for intelligence rather than use, the profits and out-of-control spending will only get worse (or higher, depending on who you are in the Industry and how you benefit from these policies — of course, it all has to be put in perspective).
George Halvorson gave an excellent overview of the underlying problems of the US Healthcare Industry. Mainly, there are no incentives for any of the main players in the healthcare system to curb spending — except for some abstract notion Americans care very little about — long term economic sustainability and the ability to compete with other countries and their growing economies.
However, George’s solutions to the healthcare problems our country faces suggest that centralization is the best method to curb costs. I strongly disagree. Controlling costs in healthcare is about decentralizing healthcare, encouraging competition (and discouraging oligopolies) and bringing it back to the community at the neighborhood level and paying for how well each physician uses intelligent communication technology to manage their dashboard of costly, chronic patients. If diabetes is the fastest growing disease in America and we get care right only about 10% of the time in diabetic patients, shouldn’t we start by mastering how to control all the diabetics that live in your neighborhood?
But of course this model doesn’t bode well for the leaders of the oligopolies that have formed in healthcare that won’t see the revenue from that $10,000 diabetic amputation.
As a side note, as I’m writing this post, I just heard a speaker say something about “disruptive technology” in healthcare being found mainly in neuroimaging.
We’ve got a long way to go folks if neuroimaging is being seen by our leaders as a disruptive technology.
by Scott MacStravic
April 14, 2008 at 2:21 am · Filed under Disease Management, Employee Health Management, Health Management
The same error in logic keeps turning up in evaluations of proactive health management ((PHM) efforts, with recent examples relative to both disease management and prevention. It is the attempt to reach an overall conclusion about PHM in general, as well as its various components, as if each is a uniform “solution” to a single “problem”. Since this is simply not the case, all such efforts are doomed to failure from the start, but manage to capture headlines when they are reached, nevertheless.
It is difficult, and probably arrogant even to attempt, to conclude whether this consistent pattern is the result of ignorance or deliberate choice, but both the media that report results in such cases, and the organizations that sponsor or conduct the analyses, are prone to the same mistake. Medicare, for example, consistently attempts and publishes conclusions about its disease management (DM) demonstration projects as if DM were one consistent, universal “treatment” to one consistent medical condition.
In a recent example, the published article on the subject reminded us of DM’s failure to save money. [R. Abelson “Medicare Finds How Hard It Is to Save Money” New York Times, Apr 7, 2008] This conclusion is based on results from eight DM suppliers who remain in a demonstration project begun in 2005, addressing different populations with different conditions by means of different interventions. Yet both Medicare and the report strive to arrive at a general conclusion that DM does not save money.
A similar conclusion was reached by another newspaper with respect to the value of prevention, in general. [D. Brown “In the Balance” Washington Post, Apr 10, 2008 (www.washingtonpost.com) ] It cited a series of studies conducted since 1986 that have come to the conclusion that prevention costs more than it saves, citing examples such as costs as high as $160,000 per life saved for men, and $240,000 for women with high levels of risk for heart attack, through conventional medical treatment therefor.
The problem with prevention, in general, is that people vary widely in their risk of actually contracting a disease, even if they have similar risk indicators. It is necessary to treat the risks in many more people than will would have turned out to be sick as a result, so it may be necessary to treat ten or a hundred more than actually deliver the desired savings of not having a disease that they would otherwise have had.
It is an unfortunate reality of the model for evaluation used in such cases that its very rigor in stipulating what is “scientific” works to limit the probability of success in what is evaluated. The best way to succeed in either DM specifically, or prevention in general, is to treat both as a “marketing” challenge – identify which people are the best prospective “customers” for using the proposed “product”, specifically in terms of their potential for success. Then, customize the intervention to match both what such prospects are most likely to “buy” and what is most likely to deliver a positive return on the investment involved.
Such a matching process, i.e. making the intervention fit the individual prospect — in terms of both likelihood of their “buying” or engaging in it, and probability of doing so realizing as much as possible of their individual potential for delivering savings – would optimize the return on investment. Unfortunately, it would also violate the “rules” of scientific evaluation from beginning to end. It would not provide a single form of the intervention, but one that varies by individual, and not to a randomly assigned group, but one composed of people who were “self-selected” to different interventions by design.
Marketers have long understood that this is the best approach to maximizing response among prospects. By at least segmenting populations by purchase potential and propensity, plus tailoring messages accordingly, the success of advertising efforts can be significantly improved. Identifying the best targets for DM or prevention makes equally good sense, in terms of promoting success. Unfortunately, it would require a far more complex application of science to demonstrate this, and too often, only a one-size-fits-all intervention, applied to a heterogeneous population, with no differences permitted is the rule for science.
Even the frequent reports of disappointing or equivocal results from DM and prevention include examples of particular interventions that do work, along with those that do not. It is only the overall picture that is reported, whereas it would make more sense to be delighted by even a few examples that work, in order to determine how to improve interventions, rather than condemn all with the same brush. If the same logic were extended to identifying which are the best prospects, and customizing interventions, the overall picture would probably be considerably more positive.
Moreover, we would learn more about what does work, instead of denying ourselves potential gains because a majority of science-restricted interventions do not. Presumably the idea is to gain success, in reducing sickness care use and expense, improving the quality of life of consumers, saving money and improving performance for employers. The aim for success should be the dominant concern when planning, implementing, and evaluating DM and preventive interventions, not scientific restrictions that reduce the chances of such success.
by Scott MacStravic
February 24, 2008 at 9:26 pm · Filed under Disease Management
In some ways, CMS in general, and Medicare in particular, have been major supporters of disease management (DM), sponsoring a large number of demonstration projects involving large numbers of DM providers and large populations of beneficiaries. They have thus gained the ability to learn more about the narrow application of DM to insured populations, narrowed still further by being Medicare beneficiaries, who tend to be seniors and to be far more likely to have multiple chronic conditions, as well as to have had them longer than is the case with employee populations.
Medicare’s limited focus is fully understandable if it sees itself as only a payer for seniors’ chronic-disease-caused medical, hospital and pharmaceutical expenses. As an arm of the federal government, however, it might also see itself, at least partly, as the source of information and insights that can guide improvements in DM, in general, and its use by DM providers, and payment by other payers, including commercial insurers and employers, as well as for consumers, who have much to gain from the best DM practices and providers.
In this latter role, CMS has failed utterly. For one thing, it has defined its role as that of determining whether or not DM works, as a generic method for a single purpose, namely saving the government money. As noted in a recent analysis of CMS’s decision in its Medicare Health Support (MHS) demonstration, the only reports on this program issued by CMS have simply mentioned that it failed to achieve the financial goals set for the program. [“MHS in Jeopardy: Part 1” Health Leaders Media Audio Feature Feb 20, 2008]
While this program has been running for 2-1/2 years, the only information that CMS has shared was that in a lengthy preliminary report on six months’ experience, which indicated that financial results at that point had not met CMS expectations. This was an overall conclusion, however, and based on the financial results alone, rather than reflecting quality of care and patient satisfaction improvements, that were supposed to also be included in its criteria for success. When CMS recently announced that the entire project will be terminated after “Phase 1”, with no “Phase 2” envisioned, it supplied no additional data on achievements in the most recent two years, nor about anything about the specific performance dimensions not achieved, merely that : “…Phase 1 of the program is not meeting statutory requirements.” [“Fact Sheet: Completion of Phase I of Medicare Health Support Program” Medicare/CMS]
It did include the statement that: “The CMS will determine whether to expand the pilot into Phase II if the results of the independent evaluation indicate that any {emphasis mine} of the programs (or program components) meet the conditions for expansion as specified in statute.” In other words, apparently neither CMS nor anyone else knows yet whether any of the individual DM programs and providers in the project succeeded, in which if any of the success dimensions, nor is there any information on what any of them achieved in the two years since the initial report came out. But CMS ended the project, nevertheless.
Considering the September announcement by the CMS Acting Administrator Kevin Weems promising a higher level of transparency and accountability at his agency, this is hardly an example of either from most observers’ perspective. [J. Young “New Medicare Chief Promises Accountability” TheHill.com Sep 12, 2007)] 68,000 beneficiaries will lose their support between July and December, with no explanation other than the terse “not meeting statutory requirements” comment, an no data backing up the comment at all, other than the first six months’ report.
On balance, i.e. over all eight MHS provider organizations, the six-month report was not positive, as far as financial savings were concerned. While savings were achieved, in terms of lower monthly costs for program participants vs. non-participant controls, when the management fee charged by the providers were included, none of them had achieved the 5% savings minimum set by CMS. In fact, two of them ended up adding to the total costs that Medicare had to pay for participants.
On the other hand, six of the eight had achieved net savings after their management fees were counted. One, in fact, saved a net 4%, though only two others achieved over 1%, namely 1.1% and 1.5%, while the others achieved only a fraction of 1%. This may be the “statutory requirement” that caused the negative decision about Phase I, though without data on the subsequent two years, it is impossible to tell whether similar results were achieved in the longer period. [N. McCall, et al. “Evaluation of Phase I of Medicare Health Support (Formerly Voluntary Chronic Care Improvement) Pilot Program Under Traditional Fee-for-Service Medicare” CMS/Medicare]
The MHS example is but one in a series of federal reports that have concluded that DM doesn’t work, where all have treated DM as if it were a single intervention form, rather than the widely varying forms that are, in reality, used by different suppliers thereof. There are far more than eight DM providers involved in the study, so the results only apply to their results, not to what others have achieved, or even what these same providers have achieved with other populations, payers, and requirements.
As one of the people speaking in the Health Leaders Media audio feature cited above argued, this is tantamount to the federal government testing a selection of eight drugs used for different patients and conditions, and making an overall judgment about the efficacy of all drugs used for all patients for all conditions. It is an absurd attempt on the face, and one that fails utterly to fulfill government’s role as a source of information that can help anyone.
If CMS had cited the data that caused them to reach their negative conclusion, and it showed that none of the eight providers involved had achieved any of the desired results, that would be one thing, but even then limited to these eight and this unique situation. Yet it adds to the already significant number of previous reports that condemn DM without mentioning the cases where it has worked even for CMS, among the different participants in its demonstrations.
Whether CMS has chosen to be an enemy of DM and its providers is unclear, but it is certainly behaving and reporting in a way that damages the hopes of chronic disease patients and DM providers, while apparently learning nothing useful in the process, for themselves, patients or providers. Shame on CMS!
by Scott MacStravic
February 3, 2008 at 10:44 pm · Filed under Insurance, Employer CEOs, Disease Management, Value-based health care, Health Management
Though most of Walter McClure’s early discussions about buying right, as well as most discussions of value-based purchasing (VBP) in general, are devoted to sickness care, it seems to me that its greatest application may end up being to health care, in its real meaning. While consumers are bound to be confounded by emotion and pressed for time when making most sickness care decisions, they can afford to both take the time and be more “rational” about choices regarding where they will look for long-range health management services.
Of course, in most cases, at present, at least, their choices are limited or dictated by employer or insurance plan sponsors of health and disease management (HDM) services, since these sponsors pay the costs thereof. Sponsors have the same challenges as do consumers when it comes to identifying the best choices for HDM services – there is very little information available, and none that I know of from any objective comparison source, on which services and providers are best.
Moreover, the “prices” that HDM providers charge vary widely, not merely in amount, but by the type of service provided, the size of the population served, and the basis used for pricing. Most HDM suppliers charge on either a per population basis, or per participant in particular HDM interventions. Such prices cannot be compared at all, unless predictions are made of the numbers of the people in a given population who will participate. Moreover, a few HDM suppliers offer guarantees or risk/reward contracts that override the per population or per participant fees they charge, where the only way the final costs and results can be determined is after results happen and are evaluated.
For consumers who buy their own HDM services, prices tend to vary widely, and again based on different foundations entirely. Many simple, Internet-based interventions are free, or cost no more than $5 per month or so, though some cost as much as $20 per month. Others are part of pre-priced packages, which may run hundreds, even thousands of dollars depending on their length, their intensity, and which health professionals offer them. Still others include HDM services as a major element of their annual retainer for “concierge/boutique” medicine, which may range from a few hundred dollars a year, through a few thousand, and up to as much as $100,000 for the very wealthy.
But there is one great advantage to HDM providers – they are already measuring and publishing their results, not merely their qualifications and care processes, which often comprise the main “quality” information offered by sickness care providers. These results are heavily canted toward payers’ interests in saving money, which may not be material to consumers, at all, but they at least address payers’ main concerns. This often includes quality measures such as participant satisfaction and perceived benefits, including self-rated confidence in their ability to manage their health status, risks, or diseases.
There is also the built-in advantage for consumers that they can take all the time they need to make choices about participating in HDM initiatives, in most cases, since there is rarely an urgent or emergent need involved. If people are screened at an employer health fair and learns that they have an emergent level of blood pressure, for example, which may happen, the usual response is to refer them immediately to their personal physician, or to a physician who can see them right away, rather than wait for an HDM program to begin.
Moreover, insurers and employees often offer useful information on the HDM providers they contract with, together with not merely free participation, but often incentives to enroll, as well. The main limitation in applying VBP to consumer decisions lies in the absence of any national body that makes comparisons across different options. We know that Medicare has found many, indeed most of the disease management efforts by specialized suppliers wanting, but these efforts apply to the narrow domain of chronic diseases, rather than the broad domain of health.
Moreover, they do not include what may be the most important outcomes criteria for employers and the health plans that serve them – the total economic impact of managing disease. Medicare deals solely with the costs of healthcare as its measure of economic impact, while employers and their health plans are looking at workers compensation and disability costs as well. And more important, they are looking at absenteeism, presenteeism, and overall productivity as well as performance impacts of HDM. These are often two to five times greater than healthcare costs, alone, and therefore much more likely to be appreciated and continually sponsored by their sponsors.
If employers and commercial insurers clamor loudly enough for, and especially if they offer to financially sponsor a national body to conduct the same kinds of systematic analysis of HDM programs and providers as has been called for with sickness care, the current lack of comparative data may be overcome, gradually at least. Already, Regence Blue Shield in the Pacific Northwest is collecting and publishing information on clinics and large medical groups in terms of their ability to manage the diseases of their patients, which covers at least chronic condition management performance.
If employers and payers expand or coordinate their efforts, measures of success and prices on HDM services may actually become a reality before similar information on sickness care. Physicians and hospitals have shown great reluctance to contribute to or accept comparative ratings by any third party, disagreeing on complex and confounding quality criteria. The criteria for HDM interventions are generally already agreed upon, by both providers and their customers, so agreeing on a set of measures should not take nearly as long.
Moreover, if payers demand comparative information from all HDM providers, in some standard format and understandable mode of reporting, providers will have little choice but to deliver it. No provider will want to be shut out of consideration for having failed to provide the information called for by their prospects and current clients. And as long as this information includes measures of results for consumers who participate in HDM programs, which it will have to in order to achieve competitive levels of participation among populations at risk, consumers will be well informed as well.
We are not even at the beginning with respect to comparison websites or “report cards” in HDM as is true with sickness care. But once we get started, we are likely to find that it take far less time and effort to create the comparative databases and mechanisms needed for VBP, to the benefit of payers, consumers, and all good performers of HDM services.
by Scott MacStravic
December 14, 2007 at 2:46 pm · Filed under Insurance, Employer CEOs, Disease Management, Research, Health Management
In managing the health of populations, there is emerging a growing consensus that managing diseases is only part of the “solution” for controlling healthcare costs. Managing the population’s overall health and reducing their risks can be far more rewarding in the long run at least. Moreover, when the population is a workforce, managing not merely risks but “productivity/performance-impairment factors” can end up saving employers far more in labor costs than they can realize in healthcare costs alone, when employee health management (EHM) is provided.
When addressing health risks, there is a substantial body of research that suggests that by reducing the sheer number of risks that members of the population of the population have can significantly reduce both healthcare costs and productivity/performance impairment. I have 33 articles in my EHM files on the subject that report either the cost impact of the numbers of risks or the benefit of reducing them, sometimes both. And I have at least three examples, all from the same source, reflecting the productivity impairment levels associated with both numbers of health risks and numbers of diseases affecting employee populations, and reporting both the costs of having them and the benefits of reducing them.
Unfortunately, there is little uniformity in: 1) deciding what health risks will be included in research; 2) choosing how many risks constitute a “low”, “medium” and “high” risk status; or 3) choosing which and how many costs of risks will be included in research. As a result, there are limited opportunities to compare results from different studies, or determine anything like average results across multiple studies.
For example, dollar costs for various individual risks, and for the numbers of risks affecting individuals, vary from a few hundred to one or two thousand dollars per person affected per year, in healthcare costs alone. In general, the costs per risk added, when only the numbers of risks were counted, usually range in the one to a few hundred dollars each, though the addition of risks tends to increase total costs per person affected more when the numbers are above five or six than when only zero to three.
The vast majority of “risks” measured have been conditions or behaviors that increase the probability of a particular chronic disease, e.g. high blood pressure and cholesterol as risks for heart disease, stroke, etc. and high blood sugar as a risk for diabetes. Obesity is a risk factor for all these diseases, as well as for arthritis and other diseases. But as a growing number of health risk assessments focus on employee populations, vs. insured lives, and on labor costs and contributions, vs. just healthcare costs, a whole new set of “productivity impairment factors” and costs associated therewith are being measured.
Productivity impairment typically identifies conditions such as high stress, emotional disorders, lack of fitness, and obesity as impairment factors, not merely disease risks. And smoking, for example, as well as poor nutrition and lack of physical activity, are behaviors that are linked to reduced productivity. Smoking, for example, is linked to the immediate loss of work time as smokers take frequent smoke breaks away from their work station and often outside their workplace, entirely. Stress is also a major impairment factor, as are emotional disorders, such as depressed and anxious feelings (mostly undiagnosed and untreated).
The costs of individual risks or impairment factors are almost impossible to determine, because people almost always have more than one affecting them simultaneously, and only the total effect of all of them, on health care costs or productivity, can be calculated or estimated. For this reason, counting the costs or impairment across a population by the number of risks is helpful, since every individual has just one number of risks, so multiple counting of separate risks is avoided.
The impact of the number of risks on both expenses and productivity can be calculated quite easily, though these are often counted separately, with employers or insurers counting expenses, based on claims costs, and EHM providers estimating impairment, based on employee self-reporting. After all, employees would be very reluctant to report their impairment to their employer, lest they be penalized for it. By contrast, such reports are kept confidential relative to individuals, with only overall impairment totals and improvements gauged by providers.
One EHM provider that has reported the productivity impairment of over 200,000 employees in its database is HealthMedia® ,Inc. Ann Arbor, Michigan. It does not have the healthcare/disability/WC costs of these employees, however, so its data understate the total costs of the chronic diseases and risk conditions, as well as risk behaviors of those in its database. And while it reports the total impairment costs associated with individual diseases and risks, it also has reported the amount of impairment linked to specific numbers of each, avoiding double counting except among those who have both chronic conditions and other impairment factors.
Its most recent report notes that the productivity impairment linked to different numbers of the seven impairment conditions analyzed ranged from an average of 6.03% for the 14.01% of the population who had less than three risks, to 8.33% for the 26.47% with three, and 11.53% for the 59.53% with four or more. Across the seven chronic risk and disease conditions analyzed, impairment started at a low of 8.2% among those with no chronic condition (they would still have impairment factors, of course), then 10.19% for those with one, 12.4% for those with two, 12.10% for those with three, 17.95% for those with four, 19.95% for those with five, and 24.63% for those with six (though only 0.22% of those in the database, roughly 440 employees, had six).
This certainly suggests that chronic risk and disease conditions are more significant impairment factors for those who are affected, but with only a minority of employees having even one chronic condition, and an even smaller minority (6.25%) having three or more, the impact of the two kinds of impairment factors was far greater for numbers of health risks, since the entire population was included in that count. The total impairment for all employees based on their impairment factors was 9.66% for example, while that linked to chronic conditions alone was only 8.51%, when the effect was spread over the entire population, since only a minority had one or more. But the average impairment of those who had at least one condition was 11.73%, where the average for those who had at least one impairment factor was only 9.50%.
Individual risk factors varied widely in their individual “effects”, i.e. the amount of impairment reported by those who reported having them, along with any others they also had. Impairment due to coronary disease, asthma and diabetes were all in the 14-15% range, while high blood pressure and cholesterol, were 11-12%, and congestive heart failure was 20.71%. Those who were obese reported 11.3% impairment, while those extremely obese reported 16.35%. Smokers were 11.86% impaired, compared to non-smokers’ 9.35%. Those troubled by high stress “fairly often” reported impairment of 14.01%, while those troubled “very often” reported 19.74%.
Among workers who reported being troubled by depressed feelings “sometimes”, impairment was 13.63%; if “occasionally”, 17.99%; if most of the time 22.78%. Even those who did not answer this question, 7.82% of the population had impairment of 19.02%.Those who slept six hours per night or less had impairment of 12.03%, while those who slept nine or more hours, and those who failed to answer the question averaged 12.23%. Those with low nutrition scores ranted from 10.32% to 18.20% as their scores got worse, while those with less than recommended levels of physical activity ranted from as little as 6% to as much as 13.54%.
Because HealthMedia only reported the number of impairment factors across a range of zero to four+, there is no real way to determine how much the addition of a single risk factor added to the impairment linked to employees with different numbers. The impairment linked to number of factors did not topped the overall average impairment of the entire population of 9.56% until the number reached four, and while impairment most likely increased from there up to seven, if anyone had all seven, the only thing reported was that with four or more such factors, average impairment was 11.53%. In any case, the key to assessing the problems for investment decisions is which impairment factors or chronic conditions, once managed, yielded the greatest overall impact.
The sheer number of conditions and factors offers some insights for use in planning EHM investments. But it is the prevalence of particular conditions and factors, together with the predicted improvement potential of each, in terms of reduced healthcare, workers compensation, and disability expenses, plus improved productivity, which should guide investment decisions. Neither the number nor the type of risks is an accurate or reliable basis for fully understanding the problem or predicting potential gains, nor even the two together.
by Scott MacStravic
December 9, 2007 at 9:27 pm · Filed under Hospital and Health System CEOs, Disease Management
The evidence on disease management (DM) at least the “scientific” evidence involving controlled trials and rigorous analysis, has been consistently discouraging. In the latest example, involving a meta-analysis of 317 separate studies, the authors concluded that while the evidence consistently indicated positive effects on care process quality, there was no conclusive support for either improving health outcomes or saving money, once the costs of DM, itself, was included. There was evidence that DM for congestive heart failure reduced hospitalizations among participants, but increased outpatient and prescription drug care costs for depression participants. [S. Mattke, et al. “Evidence for the Effect of Disease Management: Is $1 Billion a Year a Good Investment?” American Journal of Managed Care Dec 2007 670-676]
Like essentially all its preceding examples of rigorous reviews of DM effects, this one dealt only with medical/hospital care costs, not with the impact on other “indirect” costs of DM participants, for example, the known productivity and performance impairment costs for patients with chronic conditions. Nor did it examine DM intervention models that tended toward the low end of costs, since most DM programs tend to be at the moderate to high end, i.e. involve personal phone coaching, or office, even home visit interventions. By examining only higher-cost “treatments” and medical/hospital cost effects, studies have far less probability of demonstrating net cost savings.
By contrast, there seems to be ample credible evidence, though rarely meeting rigorous scientific standards, often because of the lack of randomized control trials, that DM works for employers. Random trials are unlikely to be used, since in order for employers to invest in DM, they must have high confidence already that it works, and when they do, it makes more sense to involve as many participants as possible in the DM intervention, rather than randomly assign as many as half of those eligible to a control group of “usual” or no treatment. But chances are, employers’ confidence in DM is based more on the fact that chronic diseases are major causes of productivity and performance impairment, which can add greatly to overall labor costs rather than looking solely at medical care costs.
For example, one DM supplier, HealthMedia, Inc. in Ann Arbor, Michigan, has reported the degree of productivity impairment linked to seven chronic diseases (counting both hypertension and high cholesterol as “diseases”). Employees in its over 200,000-employee database with one or more of these chronic conditions reported significant levels of impairment, combining absences and reduced performance at work. There is no way, however, to conclude that such impairment was entirely due to these conditions, since employees tend to have multiple causes of impairment unrelated to the conditions.
The specific amounts of impairment affecting those who had chronic conditions included:
- Diabetes – 4.19% of employees; 14.7% impairment
- Hypertension – 17.75%…………..11.8%
- High Cholesterol – 17.62%….. ….11.2%
- Asthma – 8.9%………………………9.2%
- Past Heart Attack – 0.73%………..14.0%
- Coronary Artery Disease – 0.65%..14.1%
- Congestive Heart Failure – 0.22%..19.4%
The overall impact of these conditions, measured based on the impairment among the 44% of employees who had one or more such condition, amounted to 11.6% average impairment. This reflects the fact that the largest numbers of employees who had one or more of these conditions had hypertension, high cholesterol, or asthma, all of which had relatively low associated impairment, while those with higher impairment included much smaller numbers of employees. The 11.6% impairment would represent a “cost” in terms of lost productivity, of at least 11.6% x $50,000 = $5800 per year for each employee affected, counting only the “wasted” compensation paid to each where the average annual compensation is $50,000 per year.
The true costs of lost productivity seems likely to be higher, as far as employers are concerned. For one thing, there is what is called the “multiplier” effect of individual employees’ absence or impairment on the team, unit or department where each works. This can range from as low as 1.00 (no effect) for completely and easily replaceable workers such as short-order cooks, to as high as 11.4 (equal to the loss of 11.4 employees) for construction engineers. The average multiplier across all workforces has been estimated as 1.28 to 1.35, though each employer should base its multiplier on the kinds of workers each has, e.g. higher for “knowledge” workers and key professionals. [“Multiplier Effect: The Financial Consequences of Worker Absences” Knowledge@Wharton, Dec 14, 2005, and S. Nicholson, et al. “How to Present The Business Case for Health Care Quality to Employers” Applied Health Economics and Health Policy 4:4 2005 209-218]
Moreover, the value of workers to employers, the contribution each makes to the employer’s financial performance, is sure to be greater than the annual compensation for each. One study found that the knowledge value alone of professionals at pharmaceutical firms ranged from as low as 1.03 to as high as 7.28 times as great as the average annual compensation paid, and averaged 2.20. [P. Strassman “How Much Is an Employee Worth? Microsoft.com/business/peopleready Jan 14, 2006]
If the average impairment per worker with a chronic condition were as much as $5800 in compensation, times an average multiplier of 1.315 (halfway between 1.28 and 1.35), times an average value of 2.2x the compensation level, the total economic loss to employers would be $5800 x 1.315 x 2.20 = $16,779 each for the 44% of workers who had one of the conditions analyzed. But this amount clearly includes impairment due to other factors, such as lack of sleep, emotional problems, poor diet and fitness, stress, obesity, etc. So the real measure of importance is how much is productivity impairment reduced, i.e. is productivity improved by available DM interventions.
HealthMedia reported on two different DM programs, one just for diabetes, and the other focusing on medications and lifestyle change compliance in general for the other six conditions. The results for diabetes reflected an average improvement of 2.00% per participant in DM, while that for the other six conditions reflected an improvement of 1.88% per participant. This would amount to the “recovery” of 2.00% x $50,000 = $1000 for diabetes, and 1.88% x $50,000 = $940 for the other conditions. With a prevalence of 4.19% for diabetes, the higher improvement with diabetes would not be as valuable as that for the other conditions, since the prevalence of the other conditions combined was five to ten times greater, overall. [E. Baas “Achieving and Measuring Productivity Improvement” (Slide Presentation) HealthMedia, Inc. Oct 25, 2007]
If the true value of recovered productivity were estimated using the team and value multipliers, it would amount to almost three times as much, i.e. close to $3,000 per participant. And this represents clear value, without any double counting as prevails when estimating the costs of individual conditions, when each participant is involved in one and only one DM program. It is likely that the value of recovered productivity will be far greater than the value of reduced healthcare expense alone. It might be better if future reviews of the “evidence” on DM included this value, rather than focusing solely on savings to insurers in reduced medical costs.
by Scott MacStravic
December 6, 2007 at 3:44 pm · Filed under Employer CEOs, Disease Management, Employee Health Management, Health Management
I think that the Heisenberg Uncertainty Principle is the one of the few things I remember from freshman-year physics in college, which was also my only venture into the physical sciences at that level. Heisenberg responded to a particularly hubris-full colleague who commented that if we could identify the position and velocity of every particle in the universe, we could perfectly forecast the totality of mankind’s future. Heisenberg demonstrated that it is impossible to simultaneously know both – one can be know perfectly, but the more perfectly you know one, the less perfectly you can know the other.
The same uncertainty applies to a major challenge in employee health management (EHM); identifying both the full extent of productivity impairment in an employee population due to identified impairment factors, and the amount of impairment attributable to each factor. In reality, while it is a feasible and relatively simple task to identify the amount of overall impairment in a population, it is next to impossible to identify how much of it is due to any single cause. And as a result, if EHM providers or their clients look at the sum of impairment due to a number of different factors, they will get a highly exaggerated and unrealistic picture of what the problem and potential solutions amount to.
To illustrate, I will use the reported impairment amounts and factors o copied from the webinar slides offered by one EHM supplier. [E. Baas “Achieving and Measuring Productivity Improvement” HealthMedia.com Oct 25, 2007] These slides showed breakdowns of productivity impairment found in a comprehensive health risk assessment (HRA) that includes questions enabling the estimation of productivity impairment. Overall impairment was linked to seven different impairment conditions or behaviors, plus seven different chronic diseases or risks. The total amount of impairment linked to each impairment condition amounted to 9.6% per person per year for those who completed the health risk assessment (HRA) upon which both impairment levels and impairment/health factors were based.
The single HRA gets the total information on each participant therein, and ends up with an overall impairment level characteristic of each individual, plus the variety of impairment factors and risk/disease conditions each has. Since the impairment is collected as an overall percentage, every time the effect of a given impairment factor is reported, the same impairment percentage is included for each individual based on each’s level of the impairment factor. And since each individual either does or does not have a factor such as smoking, or a different level of a variable condition such as stress or depressed feelings, the same overall percentage of impairment is reported across the total of all participants for each factor.
This is not true for the individual chronic risk/disease conditions, since the amount of impairment for each person with the condition is reported separately. But since individuals often have multiple conditions (over 20% of the population in HealthMedia’s data base had two or more), by counting the impairment linked to each such condition ends up counting the same impairment level many times for at least 20% of employees. The total impairment of the workforce is the same, regardless of what levels of impairment individual factors are linked to.
To illustrate, 44% of the total population with one or more chronic conditions contribute had an average of 11.6% impairment, while those without any chronic condition had an impairment of 8.1%. Those with at least one chronic condition contributed 44% x 11.6% = 5.1% the overall workforce impairment, while those with no such condition contributed 56% x 8.1% = 4.5%. The two combined add up to 5.1% + 4.5% = 9.6=% for the workforce, on average, as is true with all impairment factors.
For example, the amount of impairment linked to the numbers of health risks, for those with three or more risks, was slightly higher than the overall level for the workforce, i.e. 9.73%. But this was because employees who had 2 or fewer risk factors, representing 13.01% of the population in the database, had less than the average impairment levels, so when the lower impairment among these employees is included, the average for the workforce goes down to the same 9.6% found overall and with other factors.
If all the impairment due to all factors were added up, this would greatly exaggerate the actual total impairment of the workforce, which is, in reality, only 9.6%. HealthMedia also chose to deduct from the overall impairment level an amount it deemed to be “normal”, namely 6.1%, leaving a reducible overall impairment of only 3.5%. Still, for a workforce that averages annual compensation of $50,000 (probably average for firms that are investing in EHM), that represents $1750 on average per FTE. Moreover, individual employees showed impairment levels as high as 30%+overall, or 24%+ above the normal level, so potential improvements for some individuals could be as much as $12,000 each.
If the individual impairment levels above the 6.1% average were totaled, using HealthMedia’s figures, the apparent total impairment amount would add up to roughly $9500 per employee in the workforce, equal to 19%, when the real amount, after deducting the “normal” level, is actually only 3.5%, i.e. the true overall average impairment per employee of 9.6% minus the 6.1% normal amount. Since the total impairment tends to limit the total potential return from EHM interventions, only the overall average impairment amount can reasonably describe this potential, and at that, only if all impairment can be reduced to the normal level, or eliminated entirely.
Fortunately, the unavoidable double, triple and more counting of the impairment due to individual risk, disease, or impairment causes is not germane to choosing which make the best prospects for investment. While the total impairment supposedly due to each impairment factor is the same as for all other factors, and the total potential is the overall or average impairment level, 9.6% in the HealthMedia example, this is sure to be the limit to ROI for EHM investments. But it is the effectiveness of individual EHM programs in reducing the overall workforce impairment that counts, not the theoretical potential or size of individual problems.
by Scott MacStravic
October 26, 2007 at 9:31 am · Filed under Employer CEOs, Disease Management, Business of Health, Employee Health Management, Health Management
In most cases, sickness is a discrete variable, that is, people are either sick or not sick, and the purpose of sickness care is to return patients to the non-sick state, otherwise, they are not “cured”. With chronic illness, patients are not cured at all, merely “controlled”, by some combination of their own and their providers’ efforts. In other cases, such as severe trauma and catastrophic disease, patients are not cured, but “rehabilitated” to as high a level of normal, non-sick functioning as is possible given their condition.
The point is that in most cases, patients can be described as in one of two possible discrete states: sick or cured, perhaps controlled or rehabilitated. This does not mean that no further effort is required to maintain their non-sick state, or to prevent recurrences, crises, complications or worsening of their status, but when these occur, they will again be labeled as sick, and as “patients” rather than “people”.
In health management (HM) — of individuals, populations, employees, insured plan members or government plan beneficiaries, the targets for effort and attention, as well as the source of HM’s economic benefit, relate not to moving people from sick to nonstick, but along a continuum that ranges from perfectly healthy to having some risk behaviors or conditions of concern to having a somewhat controlled chronic condition. As long as HM participants are non-sick with some acute problem, crisis or complication, they are still not patients.
Providers may deem them patients, of course, as long as they are on the books as a member of a physician’s panel, for example. But from the perspective of HM providers, including physicians, of course, and of employers, insurers, and government agencies, they are people, who may also be participants in some particular HM program or special initiative. And while they are participants, as well as because they are, such people may be moving along the continuous dimension that runs from perfect health to death, or at least to sickness. This makes their health a continuous variable.
The significance of this in HM is that people may respond to HM interventions in any way from not at all to minimally, modestly, significantly, dramatically, etc. as high as it is possible to do so. Their response, in terms of participating in coaching sessions, monitoring their conditions, changing their behaviors and lifestyles, complying with medications, etc. will also fall somewhere on a similar continuum. And as a result of their responses, their health status will usually progress in the positive direction along a wide range of continuous dimensions, such as weight, blood pressure, sugar or cholesterol levels, bone density, etc.
And as a consequence of this positive health status change, participants will deliver benefit to HM sponsors, as well as to their own health/life quality, along yet another set of continuums. The immediate consequences along the sponsor “benefit” continuum may involve reduced sickness care use and expense, for commercial and government insurers, as well as employers. For employers in particular, they may also involve reduced disability and workers compensation costs, absenteeism and presenteeism. For the least myopic employers, consequences may also include improved retention, quality, customer satisfaction, market share and revenue, as well as profits, also continuous variables.
When the effects of HM investments are continuous variables, it is usually impossible to say that any particular HM participant has “succeeded”, since this requires a discrete distinction between successful and unsuccessful participants. HM providers or sponsors may select some arbitrary (but usually not capricious) point along the continuums of response and deem that sufficient to call it a “success”, but this does not mean that participants who made progress short of success did not deliver any benefit. In fact, it will often be the case that a participant who started of at a lower level of the continuum of behaviors or conditions but makes a dramatic improvement short of success will yield far greater benefit than another who was only a bit short of the success level to begin with and improved only to that level.
So while we can talk about and measure “success rates” for purposes of counting successes, and perhaps paying success incentives, we should not ignore that there may be a continuous degree of positive change and benefit across almost all participants, perhaps even every one of them. Even participants who do nothing more than take the health risk assessment (HRA) or screenings and get action-oriented feedback, may well improve along some health dimension and as a result deliver some benefit, despite never participating in a particular HM initiative. Similarly, others may participate actively, alter their lifestyles, improve their health, etc. but not deliver any measurable benefit to the sponsor.
For these reasons, rather than concentrate on the success rate, i.e. some percentage of participants who achieve some arbitrary point on a given continuum, the most accurate way to describe the results of an HM intervention and overall program is to calculate the average benefit delivered by everyone who can be counted as having participated at all. Separate calculations can be made for those who did no more than take the HRA or screening tests vs. those who participated but did not complete the initiative vs. those who completed, etc. in order to track how much each degree of participation added to the average benefit. But the total value of the HM initiative or program should reflect the total benefit across everyone who participated in some meaningful way.
If there are no added costs for participation, per se, e.g. if the HM provider charges on a per population basis, and the sponsor does not offer incentives that have to be paid to participants, it will make no difference how many there are in terms of the costs to sponsors, though it may add to costs for providers. But when there are costs or charges incurred specifically for participants, the evaluation of results should probably differentiate degree of participation, to see if those who merely take an HRA, or participate for a few weeks deliver enough benefit to justify the costs. If not, then further investments in persuading or offering incentives for those who complete or make some defined amount of change may be needed to achieve an optimal result.
Fortunately, there is always available, though often difficult to implement, a simple “solution” that will automatically take care of most problems related to the fact that HM deals in continuous variables. If the employer adopts a pay-for-performance system for compensating employees, all those so compensated will automatically have their health and productivity/performance levels measured, and rewarded as they improve. If such improvements are due to health improvements, this should show up in the analysis of both how much participants have improved their health in order to achieve better performance. But as long as the productivity/performance dollar value of changes made by employees is measured and rewarded, the employer will be able to determine immediately whether investments have paid off.
Of course, P4P systems usually add to employers’ costs, whether or not they are tied to HM interventions. But they rarely increase costs as much as they increase the value that employers gain. For example, when a windshield repair firm switched to P4P vs. hourly wages, it found productivity increased by 44% in the first year of the new system, while overall employee compensation increased by only 10%. Employers have always known how to ensure that the firm gains an adequate if not lion’s share of any increased value that employees deliver. And employees have usually been satisfied with a fair share.
While it serves some internal measurement, planning and evaluation purposes to identify discrete points along the continuums that reflect HM results, it is the average of the continuous benefit dimension for sponsors, and the individual personal benefit for each participant that ultimately makes the most difference. Both the average sponsor benefit and the individual participant benefits should be the dominant focus of HM planning, management, and evaluation, reflecting the underlying continuous variables that represent the reality of HM effects. A few discrete “fictions” may be useful, but the continuous nature of HM’s effects should always be recognized and reflected in its use.
by Nick Jacobs
October 23, 2007 at 7:45 pm · Filed under Policy Makers, Prevention and Health Promotion, Disease Management, Population health management
Let me open this blog with a disclaimer. It is not meant to be a criticism of freedom of choice or belief. It is merely an observation of a reality. What has caused this reality may in fact have been the phenomenal pressures applied across the world by the marketing machines of our internationally based tobacco and alcohol companies, or it may be that those present have determined that it is better to live life at its fullest for as long as we have, enjoy every bite of the apple, and deal with the reality of transition when the time presents itself.
My last four days were spent at a world conference on cardiology where the work done by our research institute’s cardiac team on the impact of behavioral modification on this disease was our presented topic. Our research revolves around diet, exercise, stress management and group support, and the results observed from our patients have been nothing less than remarkable.
It is fair to say, however, that, upon observing the actions and choices of those present my heart sank. The secret of life appeared to be firmly seated in the minds of at least 40 percent of those in attendance that tobacco, alcohol, heavy fats and little exercise are the keys to happiness.
Don’t get me wrong, it was a wonderful event where my international neighbors treated us with respect and courtesy throughout the four days of the conference, but getting in and out of each session without walking through a blue cloud of smoke, without ingesting blocks of high fat foods and free from huge quantities of alcohol consumption would have been only a dream for most.
The lounge areas of the conference were filled with cigarette stoking physicians, pharmaceutical and medical supply representatives and staff. The dinners all included well used ashtrays, plenty of cocktails and the X files list of banned foods. As a vegetarian, it was almost impossible to get through even the opening courses of a meal without unbelievable scrutiny as to my personal sanity.
That being said, the content of their presentations were as clear at this international conference as they are anywhere, i.e., the following things are very bad for heart disease: high fat foods, stress, cigarette smoking, lack of exercise and, of course, poor genes.
What then is the problem? Denial? The high pressure life styles of these life saving physicians, cultural considerations, a laissez faire attitude toward the Boogie Man or just another version of man’s on going stupidity and ignorance toward what appears to be very clear evidence?
Maybe cardiologist know something that the rest of us don’t know. Maybe they know that life is finite, that health is finite, that, like the saying I once saw on a tee shirt being distributed by a cemetery: Eat right, exercise, manage your stress and you’re still going to die.
by Scott MacStravic
October 18, 2007 at 3:38 pm · Filed under Insurance, Employer CEOs, Policy Makers, Disease Management, Employee Health Management, Health Management
Once General Motors succeeded in shifting its enormous retiree health-care liabilities to a United Auto Workers-managed “voluntary employee benefit association” (VEBA), it was probably inevitable that other companies with unionized workforces would begin doing the same. In addition to other auto manufacturers, both AT&T and Verizon communications companies are apparently looking into the same possibility. [J. Gree “AT&T, Verizon May Shift to Union-Run Health Funds” Tennessean.com, Oct 17, 2007]
In many ways, it may make sense for unions to take on the responsibility for employee, as well as retiree health, as employers seek to get out of the health insurance business altogether. Workers in unionized industries often end up working for employers whose workforce is in the same union when they change jobs, so their health insurance would become as “portable” as many health reform experts have suggested as an essential reform for the misnamed “healthcare system”. And it is certainly possible, if not likely, that unionized workers would trust their unions more than their employers when it comes to efforts to manage their health.
Regardless of whether unions or employers are responsible for paying the bill, it makes sense for both to invest in proven methods to reduce the incidence and prevalence of risk behaviors and conditions, as well as acute and chronic diseases and injuries, in order to protect the workforce, as well as control their expenses. Of course, initiating employee health management (EHM) efforts at the retiree stage is waiting until the last minute, since by then, age as well as a lifetime of unhealthy habits will have made the major challenge one of managing chronic diseases.
Judging by Medicare’s experience, “managing” older people with already established chronic diseases, often more than one at that, is one of the most expensive and least cost-effective investments in the general domain of population health management. Preventing people from adopting unhealthy habits and contracting risk conditions or “pre-diseases” generally works much better. And since this earlier option also increases worker productivity, performance, and value to the employer, it should also increase union’s success in obtaining higher wages for their membership.
One risk in shifting the burden from employers to unions is the same as that involved in shifting it from employers to governments, namely that employers may not recognize or value the opportunity to promote employee health and reduce their disease risks, once they are “absolved” of the responsibility to pay for their sickness care. But since most large employers already recognize and are doing something to improve employee health because of its impact on workforce productivity, performance and value, this may not be such a large risk.
In countries where the government has the overall responsibility for paying costs of sickness care for employees and all citizens, employers have begun taking on added responsibility for both private insurance for sickness care, and internal EHM programs for their employees. As most US employers recognize, enabling employees to access sickness care when they wish, rather than waiting in line for providers in public insurance systems, is a major recruitment and retention “perk”. Moreover, promoting employee health, while reducing productivity and performance impairment factors, can be a highly cost-effective approach to reducing overall labor costs and even increasing revenue.
Even if employers manage to be totally free of any responsibility for employee sickness care expenditures, it will remain in their best interests to protect and promote their employees’ health. The positive economic, bottom-line impact of doing so will still be from two to five times as great as are sickness care cost savings alone. This fact may even become the basis for a more cooperative vs. adversarial relationship between management and labor as unions take on the responsibility, since unlike negotiations over wages, cooperation in EHM would be a win-win opportunity for both, as well as for employees, dependents and retirees.
Next entries »
