Which Health Policies Actually Work? We Rarely Find Out

The Rand Study showed that reduced use of health care didn't have short-term adverse health outcomes... "(except for the poorest families with relatively sicker members)." Well. Was the reduced spending worth the worse outcomes in those families? Or don't the poorer, sicker matter?

This comment may not be entirely relevant to the article, but I think it’s important within the the context of improving patient convenience and care. At one point or another, we’ve all encountered the situation where our primary care physician orders a visit with a specialist, who, of course, is at another location and requires a new set of forms to be filled out. I would like to see the U.S. healthcare system move to an infrastructure model where a single facility or structure contains not only general practitioners, but also physicians focusing on a variety of specialties (optometrist, cardiac specialist, etc.), including all of the specialized equipment. A patient—based on a preventative model— could get everything checked out in one day, with no additional travel or paperwork. Yes, it might be a challenge to arrange, but the long term benefits would easily outweigh the short term organizational problems.

In order to evaluate policies/programs, data is needed. In the US, there is no centralized national patient database making comparisons of metrics very difficult. Even with Medicaid/Medicare, the patient information database and the claims database are completely separate. So if one wanted to compare, let's say, which program had better success treating diabetes by using the lab result HgbA1c, the impartial investigator would first have to get a list of patient names, then call those patients to find out what doctor they saw and then try to get the patient to sign a release to get copies of their blood test results and then send those releases to the doctors' offices and then hope that they will cooperate in order to obtain that information. So our first problem is a data problem. A way to start moving us in the direction of being able to optimize policies/programs by way of data collection in the healthcare space would be to first, start out with everyone having a national healthcare card/number that stays with them for life which has an automatic opt-in mechanism to collect laboratory/testing data/vaccine information; in order to be exempt from this data collection, you'd have to opt-out.

'Sick around the world", a doc on PBS in 2009 shows how nations around the world deal with healthcare. America has always had a problem with taking advice from other countries. Arrogance plays a part which I don't understand. A solution could be close and we turn away. Short sighted policies.

Set a voucher payment price (likely by region to account for cost differentials) that Medicare/Medicaid will always accept as payment in full, and then patients choose what to do with it in terms of buying health care. Why demand one-size-fits-all when we know that's never the case.

This kind of musing about health care policy is worthless when, in fact hospital systems are the drivers of US health policy, and insurers are the enablers. And using Medicaid as any kind of test situation is flawed, as it’s a limited coverage administered differently and randomly by states. As for Oregon, it’s the state that spent $200 million to launch an ACA online insurance market that never worked.

I remember reading in another article how private insurers rarely pursue action against insurance fraud. The cost is just passed on to their customer's. My surgeon would have to be counterbalanced by other income to make up for the minimal paid by Medicare for my surgery. When data as regards weather is being altered to conform to someone's ego; there's little chance complex Medical data will conform to actual results.

There is also hardly any evidence based research done on which features or functions in electronic medical records benefit patient care nor which risks they might increase.

Controlled trials are useful to test a hypothesis, however NIH sponsored trials frequently are presented to show statisitical benefit but not practical societal benefit. Consequently even medical practice based on clinical trial results may fail to have either sustained medical benefits or even more frequently fail to be cost effective. Although controlled trials may be good to test hypothetical outcomes in Medicare health policies, what is needed as in all of medicine is better long-term benefit data for any accepted standard practices.

O, okay. I will now stop feeling like a complete bonehead for struggling to endure/understand my (crash) course in research concepts, which was of course required for the Master of Science I pursued and earned. Sheesh.

What an absurd idea, to make decisions based on evidence!

Agreed, there needs to be more rigor in both research, and application of results. The US Preventive Care Task Force does assess evidence based practices and issues guidelines on what is of therapeutic value, and what is not. Health insurers use these as guidelines on what to pay for. The Dartmouth Atlas of Health Care shows the wide variation in practices means the care we get is often defined by where and when the practitioner went to school. But take another look at the Oregon Medicaid situation, a lot more work has been done on health metrics since that initial 2 year study period.

I wonder if this is the case with country like US , where is the evidence for developing countries? Is it because there is a limited advocacy to take evidence based decisions or limited capacities in countries to conduct evaluation of policies. Once the policies are made, there is hardly any chance to alter it through new evidences. Countries like India where policies are made on populist sentiments of people with no authority to understand the impact of certain policies over other policies. We should empower communities so that they can ask government to take rational decisions taking inputs from rigorous studies on health policy analysis. We should have a dedicated unit in Health Ministries to evaluate policies and guide us to take corrective action to achieve desired impact.

There were "no short-term adverse health outcomes" when people who had to pay more for health care decided not to see a doctor. How about long-term adverse outcomes? Cancer can take a while to make you so ill that you realize you must skip some other bills and forgo some food. So can diabetes. So can the shortness of breath and fatigue that will eventually surprise you with a heart attack. So can the symptoms of obstructive sleep apnea, if you ignore them and your brain spends months and years worth of nights with insufficient oxygen. Let's not be quick to decide that having health care readily available for all is not a good idea. Instead, please compare statistics on the health of people in a first world country such as Canada with the health of people here in the US. (I can't decide whether we're tumbling into third world status or being kicked into it by GOP politicians.)

As the administrator of several hospital-based addiction treatment programs, I had to attempt evaluation of our patients' post-treatment status at prescribed intervals, such as 6 months and one year. I had no funding to do so, but evaluation was mandated by JCAHO, whose standards we had to meet. We made some effort to call these patients, but no one in our department really had the mathematical expertise to evaluate outcomes, nor was the number of patients we reached sufficient to draw any real conclusions. Furthermore, the hospital considered our program a profit center. It would be untoward to have numbers that did not prove our efficacy. We offset such hospital programs as cancer treatment, which are loss centers. To balance the non-profit budget, we had to stay full and profitable. I think we did a good job of treatment. I know I had an outstanding staff. I just wish we could have proved all this.

While it may be impossible to determine the efficiency of a particular program without randomized controlled trials,the efficiency of more general principles can be inferred by long term observations over many different situation. The science of epidemiology is based on this. Thus it can be seen that universal, government run health care system are much more efficient than any system that relies largely on private insurance. This is based on decades of data over many very different countries with different universal government systems compared with the US.

Healthcare is a human right. Period. The US has signed and ratified the Universal Convention on Human Rights, which establishes healthcare as a human right. What kind of human being denies his fellow citizens care claiming "no short term health effects"? How does Mr. Frakt sleep at night?

In his Sunday (Sept 8, 2019) column for the Times entitled: "The Age of American Despair," Ross Douthat discusses a crisis that is "killing tens of thousands of Americans," annually. Termed "deaths of despair," he states that there has been a "sudden rise in deaths from suicide, alcohol, and drug abuse since the turn of the millennium." He adds: "a doubling from 22.7 deaths of despair per 100,000 Americans in 2000 to 45.8 per 100,000 in 2017" highlights how critical the crisis is. [325 million people in US x (45.8 deaths/100,000 US people) = 3250 x 45.8 = 148,850 deaths annually] Combining the psychological nature of many elements of this crisis, and the huge costs of treatment, with the huge subjectivity that funds have been allocated on past mental health projects, any improvements in the objective evaluations of these policies and their projects are a good investment. Many of these psychological treatments will probably struggle with whether they can be replicated or not, and to allow the more overarching policies from which they spring to ignore random selection and assignment protocols, seems almost obscene, in a way. [09/09/2019 M 5:01 pm Greenville NC]

I have been a manuscript reviewer and editor for Swedish medical researchers who must write in English, researchers who often write pretty well but not well enough. This gave me a window if a small one into research of the kind in focus here research evaluating particular changes in Swedish Universal Health Care policy. Even now I do this for a truly world-class researcher in a very important field I choose not to name, a field in which completely new approaches to patient care are badly needed. That researcher can write grant proposals that often point to something well known to American researchers who work with Swedish researchers in certain fields. That something comprises the databases in which the health-care life histories of all of us are kept. Every time I make use of the SE UHC system, my visit is recorded via my Personal number that begins with 6-figure date of birth. This results in the availability of a vast body of data perhaps making studies possible that are very difficult to make in the US or perhaps in a region like my New England. I end by pointing to that sentence: "Half of treatments used in clinical practice lack sound evidence." That is exactly what my world-class researcher points out so clearly in grant proposals and in the research done by "hens" (Swedish his/her) teams. Only-NeverInSweden.blogspot.com Citizen US SE

Thank you, Prof. Frakt. Those of us who teach research methods and conduct serious policy studies have agonized over the lack of rigorous evaluations of healthcare programs and policies. Today, a congressional committee is moving on a proposal to allow unique patient IDs. About time! We have frightening estimates of the number of wrong patient errors made each day--errors that cause patient harm and death. We should not base so much treasure and so many lives on guesses and wishes. We have the skills and methods to determine what works and what doesn't. We should use them. Ross Koppel, PhD, FACMI

CDC's National Center for Health Statistics has statistical apparatus which could be expanded to collect all data on all health care outcomes in the U.S., if they were given the mandate and budget to do it. Why has this not been done? Guess: The insurance and health-care industries strive mightily to keep federal regulation from affecting their income streams. Their lobbyists have leaned on Republican legislators to keep this from happening.

This was a smart, thoughtful piece. It remains staggering to me that we spend so much of our national treasure on health care, yet wish to remain rather blindingly ignorant about whether our policies work (or work as intended), and how we can improve. California has an entity (which I direct) that provides independent analysis of pending health policy legislation directly to the Legislature (but does not issue recommendations). Our work highlights available evidence, and provides a grading of the strength of such evidence (based on sample, robustness, and research design). Our work is designed to enhance policy-making and bring economics, clinical and public health expertise to bear as legislation is considered. All of our more than 150 reports are available at www.chbrp.org, and we enjoy bipartisan support and have been hard at work since 2003.

We used to have an agency (AHCPR) dedicated to figuring out what works in health care but it was killed when it produced a guideline that irritated back surgeons. Michael Lewis has a series of podcasts on the death of the referee in America. This issue is just one example.

I'm surprised Austin didn't mention the classic case of policymakers ignoring evidence: abstinence-only education. The bulk of evidence shows it simply doesn't work and yet it keeps getting funded. Ideology trumps efficacy. Another issue not raised is the difficulty of teasing out causality in real-world settings. Right now there's so change much going on at so many different levels that it's very difficult to isolate what is influencing what.

While the second half gets into some of the complexities of rigorous evaluation, defined as randomized designs, it's still presented as a panacea. There are issues with it, some of which are fairly well known and some are subtle. Fundamentally, it's a tool; not an infallible umpire. On the better known side, if you have a potential intervention that seems to be working well in a troubled situation, is it ethical to withhold treatment from the control group, some of whom think they're getting it? Or, end the trail, declare victory, and try to get everyone well. Another is the political system is not built for this. It is a lot easier to yell "Failure" on the news about a policy tried by an opponent, discouraging them from evaluating it or communicating the result, then work with them to catch and adjust an early problem. More nuanced, in a comment on another article today about Israel, I discussed the common definition of success as restored security is not sufficient if it ignores accelerating a future backlash. Designers of a randomized trial should ponder, and then, as more is learned, re-ponder that misleadlingly simple question. Then there are multiple interpretations of what the data says (the truism "data speaks for itself" is not always true); how are intangibles evaluated; can the evaluation system handle surprises, including useful ones; as well as the mentality that reality only exists if it has a number. As a tool--fine; not as a panacea, and in field after field.

My state’s expansion of Medicaid has been a complete blessing for me. I’ve gotten a digestive order conquered and found out what I thought was stubborn neck, upper back, and arm muscle pain is actually four severely deteriorated discs, which will be replaced 9/25. Next year, after recovery, I’ll be able to address lower spinal issues. Conservatives may call me a freeloader, but at 58 I’ve paid a lot of taxes since I started working 41 years ago, twice been a caretaker for a dying relative, and have gradually become unable to work full-time because of my health issues. I endured a decade without insurance. I sincerely hope I never have to face that fear again.

IMHO, every medical provider should be required to accept EVERY SINGLE medical insurance offered. Think about it: The reason they don't is because the insurance company decides how much they pay for ANY service. Why do insurance companies get to determine how much they will pay for a service? That's like dictating the price I pay for a Mercedes Benz. Listed @102,895? Hah! here's a check for $ 900. Deliver MY CAR. About the lack of "studies": How much will you wager that the insurance companies dictate what is studied and what isn't, by dictating how much graft they will send to our elected officials? Thanks to my $ 1562 monthly medical insurance premium, I've only got $5 to wager, but wth, I'll go all in.

1-The vast majority of our peer countries Health plans work. They are by and large satisfied with it and looks at ours as de facto criminal. 2-Medicaid and Medicare are variations of above for selected people in our country. Medicaid prevents poor people from being thrown out in the street. Medicare is loved even by far out conservatives. 3-The rest of America is under a de facto criminal system. Above are the basic facts. Any questions?

This opinion may be correct in its assessment of operational challenges, but it misses the political angle entirely. Unlike the other comments, I mean political not in terms of the easy graft and campaign contributions. I mean Political in terms of the research does not establish generally agreed upon outcomes. As we saw in the ACA debate, not everyone agrees that everyone should have access to a longer life expectancy. Even cost cutting is controversial- "everyone else, sure- but I want the best possible because I'm not paying for it". Until there is a set of agreed targets, there is no reason to study. If the goal is the health of every citizen, we would spend more on lead paint remediation, pollution control, and rehabilitation of low income housing- not medical care. The Oregon example is perfect. Is a lack of financial distress and improved mental health for the poor a good in an of itself? Not according to people just outside of the qualifying criteria for the program. In other words, define what efficacy looks like and then come up with ways to track against it.

I just wish there existed a place called THE ENTIRE REST OF THE WORLD, where healthcare policy included every single citizen and was completely paid for by public taxes. A place where we could see in reality a system that cost half as much as we currently pay, and yet yielded far superior health outcomes. A place where its people could try their best to cope with the tragedy of serious illness, without the additional tragedy of bankruptcy heaped on top of it. If such a place existed, we would look at our current system and realize what gullible idiots we have been for decades.

Much of medicine is based on faulty theories. The annual physical is one which, at long last, even the AMA decided is not cost-effective. It is rare to find conditions worth treating. Yes, it very occasionally happens but not often enough to inconvenience everyone else and to spend a lot of money on tests. Stats finally showed that most women do not need pap tests or even pelvic exams yearly, but old money-making habits die hard. The Bill and Melinda Gates foundation pays for separate statistics to test their grants. Sometimes the stats show the grant is successful and sometimes it needs to be changed. It is shocking that physicians don't care what works. Hey, they learned it in med school. It is more shocking that our government rarely wants to bother to check out the cost-effectiveness of policies. They do occasionally. I remember when seniors in Rochester NY could get free flu shots when others could not. That worked out so well in cost savings for hospitals that the next year it was policy. They should be doing more testing of stats to save us all a lot of money.

Interesting that employer wellness programs have no impact on employee health. I think the key to this failure is that the programs had no impact on what employees eat. Food is the key to health--the amount of sugar, saturated fat, salt, pesticides, and chemicals (as in processed meats). But one has only to observe one's fellow shoppers in stores to see that Americans are morbidly obese, barely mobile, compared to our parents' and grandparents' generations. How to get the message across that organic, low-meat, low-sugar, heavily fruit and vegetable, whole grain diets are the key to health? Some of us pay attention and know the value of good diet and exercise, most of our fellow Americans don't. Though food is probably the key to health, the increase in pesticides in yards, the air, water, and on our food is probably the second biggest cause of ill-health and birth defects. But these practices are so easy to correct, if only we cared enough to read, and act.

Conducting effective studies then getting the results into policy are extraordinarily difficult. Pilot studies are pretty much useless because their results are virtually never replicated in real world, generalised settings due to slippage, random variation, quality of participants, funding levels and much else. Politically inconvenient research results (e.g. many drug and surgical interventions are either useless or harmful; mental heath programming is pretty much never effective) won't ever find their way into policy. We've known since Engels wrote his tract on the English working class in 1844 that improving housing, income, working conditions, diet, neighbourhoods, parks and education will have a much greater impact on human health and longevity than any and all medical treatments and social work interventions, but in America there's a stubborn refusal to do much about any of those economic and social factors because they threaten markets and neo-liberal values.

I used to be a researcher for Medicaid. The truth is that even when we had findings about which policies work and save money, legislators do not care. They have an agenda and they do not want to know if their agenda is harmful to patients and/or does not save money.

The proof that public health care works better than the current American system is in the longer life spans and lower health care costs in the countries that rely on a public system: namely, almost every Western Democracy other than the US. In Germany, which has an insurance option, there is extremely rigorous oversight and regulation of insurance offerings, pharmaceutical products and prices, and health care generally. Of course, the shrinking lifespan of Americans is also due to economic and social policies that fail to support public health and well-being.

What this article misses is that we have good data on the big picture of 'what works.' The problem is our US system makes a few people rich, and they have outsize influence. What do we know? From studies comparing nations and areas within nations we know: Primary care saves money and saves lives. Barbara Starfield demonstrated this beyond any doubt back in the 1990's. Countries with nationally paid and nationally delivered health care have the lowest costs and highest quality (Britain.) Countries with nationally paid and privately delivered health care have the next lowest costs and next highest quality (Canada, Scandinavia.) Countries with strictly regulated private nonprofit insurance and private health care have much higher costs and variable quality (Switzerland, Germany.) The US, with private for-profit insurance (which also runs US Medicare) and private, for profit health care with limited regulation, has the highest costs by far and the lowest quality. (Data from Commonwealth Fund studies.) In all the OECD nations, the SUM of health care spending and social services spending is about the same percent of GNP. The US falls between the Netherlands and Norway, though they spend twice as much on social services and half as much on 'health care.' Social supports have an outsize effect on health. They get more for their money. (Again, Commonwealth Fund.) We know the big picture. Big money blocks better systems that would cost less and hurt their profits.

In a for-profit system there are many with a very strong motive to direct money into their interests. And an even stronger motive to keep that money flowing once it starts. Rigorous evaluations cost money, require an arbiter without a stake in the outcome, and may end up costing someone a lot of money if a treatment protocol or medication is found to be ineffective. That's why they are exceedingly rare.

Health policy research is potentially beneficial, but it should not distract us from adopting policies that do not require research to know whether or not they work. Currently much research is directed to studying value over volume, yet the preliminary results have been disappointing, and the response has been to call for more such studies. This is actually harmful when it deters us from adopting policy changes that would be inherently effective. Current research on alternative payment models is largely wasting resources, partly through an increase in useless administrative complexity. On the other hand, we can enact and implement policies that would automatically include everyone, sharply reduce the profound administrative waste of private insurers and the burden they place on the system, make health care affordable for everyone by financing the system with equitable progressive taxes, and improve access by removing financial barriers to care. You do not need a complex policy study to see whether or not those goals would be realized. They are inherent in a well designed, single payer model of an improved Medicare that covered everyone. Policy research is fine, but let's not walk away from commonsense solutions just because the policy community hasn't done the research yet, when no such research is required.

Bean counters often forget that the United States has been designed to constantly run up to 50 massive experiments simultaneously. When each sovereign state consider legislations we all get to see what works best. Consider the legitimate health crisis of gun violence and the mental instability at its root. Next imagine a few states passing legislation requiring mandatory gun insurance that requires the insurance companies to price the risk of harm of particular firearms and ammunition in the hands of particular individuals. The task of reviewing and monitoring the social, education, religious, cultural, financial, and health data of gun owners would spur advanced artificial intelligence software that would consider all statistically reliable human factors (that the government should never be trusted with without a subpoena). Consider even grander experiments where a few states: 1. provide all prescription medication without cost 2. nurse practitioner examinations, prescriptions, nutrition guidance, and doctor referrals are free 3. health insurance companies are required to maintain and share digitized records from all health care providers (and provide free access to public health researches and AI summaries of best care practices) 4. men have a cause of action for damages against an abortionist that destroys his unborn child without consent If the federal government imposes a 50-state solution (i.e. Obamacare) experimentation and the potential for rapid progress dies.

This is common problem with most government programs and research where funding to evaluate and publish results is often too meager and too short term. However, as the 2015 study showed, it remains true that hospitals do cross subsidize. The core issue is who pays the unrecovered costs from the underinsured and uninsured? One study, done by our nearby regional hospital/clinic offered hope. With a slight expansion of medicaid and the mandatory insurance at the beginning of the ACA, unrecovered costs were 40% lower for the 5 years after the ACA compared to the 5 years prior to the ACA. it plateaued at the new lower level in the subsequent years but has risen again with the removal of the individual mandate. Now that is some research that needs to be done.

I strongly think that using good data and studies to drive health policy is essential, but several issues undermine adoption: One, a complaint I hear from medical personnel is the crushing amount of work involved. The data collection often means hours of extra, unpaid work, and the only respite is likely to doctors, with scribes and dictation tools to handle the extra labor. The crushing amount of work leads to a hatred of the system, and in its worst cases, depression. An additional problem is the time lag between study results and practice. This is a problem that will haunt AI as well, in that knowing something is better rarely becomes the norm, but can take years for the medical profession to change policies. Even then, the nature of American health care and its political subversion is that those looking to undermine positive reform will use fear tactics and ignorance to push public opinion away from effective rationalization.

Some myths are dispelled in this story but one thing that seems to make sense and that almost every one agrees upon is that our government is not doing nearly enough to combat fraudulent use of our Medicare and Medicaid systems.

After some years in practice, it dawned upon me that we knew only isolated pieces of useful information, in a sea of vast ignorance. That hasn’t changed, in my view. The more data we collect, the more we devise concepts to connect and “understand” their implications, the greater is the complexity of biology revealed. The practical point at this time is only that a patient working with a conscientious physician is likely to do better than on her/his own. Stephen Rinsler, MD