You are currently browsing the archive for the Medical Studies category.
The Economist’s Technology Quarterly reveals some recent advances in medical technology:
The physician-patient interaction can be a strange one. Patients leave their most important possession–themselves–in the hand of strangers. Typically conservative women will bare their naked bodies to physicians. Although rare, the possibility exists for the physician to take advantage of this situation.
In the U.S. “4 percent of the disciplinary order that state medical boards issue against physicians are for sex-related offenses…Seventy two percent of females medical students and twenty nine percent of male medical sudents have been victors of patient initiated sexual behavoir.” I know of patients who slap cute medical assistants’ behind. Physicians and patients can both act inappropriately.
How do physicians establish trust with their patients?
Atul Gawande gives some examples from around the world in Better: A Surgeon’s Notes on Performance.
Gawande, Atul (2007) Better: A Surgeon’s Notes on Performance, Metropolitan Books, 288 pages.
Gawande A (2005) “Naked” NEJM, 353:645-648.
What factors predict how long we live? What are the best ways to forestall death?
The determinants of premature death are 40% behavoiral, 30% genetic, but only 10% medical care. It is important to remember that medical care and health are far from synonomous.
With the Senate passing a $700 billion Wall Street bail-out last night, the Healthcare Economist wonders who else needs a bail out. The best and brightest health bloggers have your answer. In this edition of the Health Wonk Review, we will examine six groups looking for help:
WALL STREET
HEALTH INSURERS
HEALTHCARE REFORMERS
DOCTORS [Especially those owning MRI machines]
THE UNINSURED
KIDS
INTERESTING POSTS that I couldn’t tie in with the bail out theme
The N.Y. Times Well Blog writes that 19% of men regretted having prostate surgery. What is interesting is that men who underwent a newer, less invasive, robotic surgery were four times more likely to regret the prostate surgery than those who underwent the older, more invasive “open” procedure.
Is this increase in regret due to worse outcomes? This is likely not the case. Instead, it may be that doctors are inflating the patient’s expectations of how life will be like after the robotic surgery:
“Part of the problem may be that doctors who perform robotic prostatectomies commonly cite potency rates as high as 95 percent and above among their patients, giving patients an unrealistic view of life after surgery.
But the data are highly misleading. Researchers often define potency as simply being able to achieve an erection that is “adequate” for intercourse — but for many men, that definition doesn’t capture their ongoing struggle to return to a normal sex life. Earlier this year, researchers from George Washington University and New York University used a more realistic definition of potency, showing that after surgery, fewer than half of the men studied felt their sex lives had returned to normal within a year.”
The success of a surgery may not only depend on the technical skill a physician exhibits, but also how they are able to control the patient’s expectations so that they are able to lead a satisfying life after their surgery.
On Friday I reported that the U.S. scored poorly on the Commonwealth Fund’s National Scorecard. Those in favor of universal health care are probably rejoicing. “The U.S. system is dysfunctional beyond repair and we need universal health care!”
Yesterday, the Economist reported on an article in The Lancet Oncology journal which found that the U.S. has the best five year survival probabilities for breast and prostate cancer. Score one for those against universal health care. “The American free market is always the best!”
How can this be? How can we reconcile these two results?
The Lancet Oncology article controls also for other covariates which are related to survival probabilities, but do not relate to the quality of health care. For instance, if Americans get cancer later in life than people from other countries this is taken into account since people who are older are more likely to die of almost all causes, including cancer. Further, if traffic mortalities or the homicide rate are higher in the U.S. than in other countries, this will likely decrease the probability a cancer patient survival for 5 years, but is unrelated to the quality of medical care. If Americans are more likely to be obese, this also will decrease their survival probabilities, but should not be an indictment against the health care system. For these reasons, the 5 year cancer survival probabilities are adjusted to take into account the age and death rates in the general population. After these effects are taken into account, the U.S. scores very well in terms of cancer survival.
Of course cancer survival is only one of a myriad of ways of measuring the quality of the American health care system. Further, the U.S. spends the most money on healthcare (in total and per capita) compared to any other country. While the U.S. may (or may not) be the best, it is certainly the most expensive.
According to InsideHigherEd.com (”Tomorrow’s Doctors…“) medical students are very altruistic, empathetic people…until the start medical school. The article describes the findings of a study titled “Is There Hardening of the Heart During Medical School?” in March’s Academic Medicine.
The longitudinal study finds significant decreases in “vicarious,” or emotionally driven, empathy, during the course of medical education. Significant drops happen after the first year and after the third, clinical year when “students,” the article notes, “were seeing patients they had, presumably, looked forward to helping.” (The drop at that point of first patient contact in the third year is particularly concerning, the lead author, Bruce W. Newton, said in an interview Thursday)…
The authors find, for instance, that students who choose the “core” specialties, where they see many of the same patients (i.e. internal medicine, family medicine, pediatrics, obstetrics-gynecology and psychiatry), manage to better maintain their empathy throughout medical school compared to those who choose “noncore” specialties (like radiology or surgery), where continuous contact with specific patients is limited.
[Hat tip to Marginal Revolution]
Differences in the health outcomes between white and minority patients has been well documented in the medical and economics literature. Reasons for this difference could be:
A recent paper by Emilia Simeonova tries to dig deeper into what is causing the racial mortality gap for chronic heart failure (CHF). CHF is one of the leading causes of death for the elderly and one of the major components of the racial mortality gap.
Methods
Ms. Simeonova uses a six-year panel data set from Veterans Affairs [i.e.: the VHA Medical SAS inpatient and outpatient datasets,the Beneficiary Identification Records Locator Subsystem (BIRLS) death files, the VHA Enrollment files, and the Veterans Service Support Administration (VSSA) clinic performance measures database]. The data allow the author to compare treatment within facilities rather than just between them. This is important because it is possible that blacks go to bad doctors and whites go to good doctors and this may constitute the entire mortality gap. By comparing outcomes within a clinic or within the same doctor, the author can better analyze what is causing the mortality differences.
The author calculates 3 year survival probabilities conditional on surviving two years. This should help to eliminate different CHF severity levels. In her regression, Simeonova uses patient and clinic characteristics, as well as clinic fixed effects, and time and cohort dummies. Simeonova measures doctor quality as the probability the doctor prescribes beta blockers and ACE inhibitors to patients with chronic heart failure (CHF). However, another aspect of the quality of medical care is patient compliance. Patient compliance is calculated as the number of prescriptions filled on time divided by the total number of prescriptions filled.
Results
Simeonova finds that doctor quality accounts for 5% of the CHF mortality gap and socio-economic factors account for 20% of the differences in CHF mortality. However, the vast majority of the mortality differences are due to the fact that blacks are less likely to take their medication than whites.
“I show that doctor quality significantly influences patient outcomes. While minority patients visit slightly less competent doctors, this does not explain the large gap in survival. Individual doctors are found to treat their patients similarly regardless of race. On the patient side, I demonstrate that variation in compliance triggers a racial mortality gap. Differences in patient response to treatment significantly alter survival probabilities. Considerable reductions in medical costs could be achieved by convincing patients of the importance of strictly following the therapy regimen. I estimate that targeting compliance patterns could reduce the black-white mortality gap by at least two-thirds.”
Also interesting is that the paper found that when blacks have a regular doctor, they end up seeing a lower quality doctor. Nevertheless, compliance rates and mortality decrease for blacks when they have a regular doctor despite the fact that this doctor may be of a somewhat lower quality.
The USA Today reports on the development of a shingles vaccine. According to the article, “The vaccine reduced shingles cases by 51% in people given the vaccine vs. those given the placebo. Vaccination reduced the burden of illness, a measure of pain and discomfort, by 61%.”
So why aren’t people getting this vaccine? One reason is how the vaccine is paid for. The vaccine, priced around $150 by the manufacturer, is covered by the part of Medicare that pays for prescriptions, not doctor visits. That means doctors are not automatically paid for shots given in their offices. Some send patients to pharmacies to get the shots or pick up prescription vials, adding steps that may reduce use, Oxman says. Others stock and give the vaccine, but require patients to pay upfront and seek their own reimbursement.
Last night I saw an encore presentation of an interview with Jerome Groopman on the Newshour with Jim Lehrer. I have written posts in the past about Dr. Groopman (see 16 March 2007). Dr. Groopman notes that doctors often misdiagnose patients. Physicians are often anchored to a prior belief and have trouble changing their diagnosis even when new evidence appears. Dr. Groopman advises that there are three important questions that patients should ask their physicians after they receive a diagnosis:
There is an interesting article at Forbes describing that the housing boom is not the only bubble that may need to burst. Scans per thousand insured people went from 85 to 234 in the U.S. between 1999 and 2007. Author David Whelan describes what happened to one radiologist in Connecticut after Medicare and HMOs cut scan reimbursement rates this year:
Radiologist David Gruen used to spend millions of dollars to replace his General Electric MRI and CT scanners every three years. It was money well spent because the machines were always busy. But a year ago Medicare cut the price it pays for imaging, so Gruen gets paid 15% to 50% less for each order, depending on the type of scan. Health insurers got wise, too, and started imposing a 48-hour review on imaging orders. The doctor hired clerks to battle the HMOs, but his office volume was flat last year, down from 10% growth in prior years. Gruen was forced to take a 20% salary cut. Now his Norwalk, Conn. practice is holding off on buying new machines and stretching the old machines’ life span to five years. “We really do face a crisis,” he says.
In additional, General Electric’s (disclaimer: my former employer) medical division has seen declining profits for the first time in years.
Are Medicare and HMO cuts to imaging hurting patients? The New England Journal of Medicine thinks not. There are side-effects to these scans including increased levels of radiation exposure, especially dangerous for kids. As with any test, there is the probability of a false positive (i.e., that patient does not have the disease, but the test claims they do). “A study from the National Institutes of Health found that 17% of patients getting tested for cancer had at least one false positive chest X ray over a four-year period, and 8% of women had at least one false positive ultrasound for ovarian cancer.” These figures lend some more evidence that Americans may be Overtreated (see my post on Shannon Brownlee’s book of the same name).
Forbes also finds that “a doctor who owns his own machine is four times as likely to order a scan as a doctor who doesn’t.” Financial incentives do make a difference (for more information on how physician financial incentive affect surgery rates, see my working paper “Operating on Commission“).
Is the current medical school curriculum optimal in terms of teaching doctors-to-be how to best practice medicine? DB’s Medical Rants blog says no.
“One cannot really understand most diseases if one cannot correlate the physiology. We should focus our anatomy teaching on the big issues, not the 3rd branch off the 2nd artery (hopefully you get my point.) We need not have students be able to identify dried out cadaver structures. Rather we should learn functional anatomy – e.g., the structures and causes of shoulder pain, the types of knee injuries, the correlation of DVT with clinical findings, intra-abdominal anatomy, how specific brain lesions impact patients.
To support these 2 very important courses we need to learn enough cell biology and histology to understand the cellular components of disease. We also need to know enough biochemistry to support the physiology course.”
Sounds like good advice to me.
A New Yorker article (”The Checklist“) recounts Peter Pronovost’s efforts to improve the delivery of medical care. One of his simplest ideas was to invent a 5 step checklist to reduce line infections:
Doctors are supposed to (1) wash their hands with soap, (2) clean the patientâs skin with chlorhexidine antiseptic, (3) put sterile drapes over the entire patient, (4) wear a sterile mask, hat, gown, and gloves, and (5) put a sterile dressing over the catheter site once the line is in.
All doctors know these 5 steps, but in the distraction-filled world of the I.C.U., it is very easy for the physician to forget any one of the steps. Dr. Pronovost’s checklist idea has extended to other treatment areas as well. Yet he believes that Americans are still not getting serious about treating medical care as a science.
âThe fundamental problem with the quality of American medicine is that weâve failed to view delivery of health care as a science. The tasks of medical science fall into three buckets. One is understanding disease biology. One is finding effective therapies. And one is insuring those therapies are delivered effectively. That third bucket has been almost totally ignored by research funders, government, and academia. Itâs viewed as the art of medicine. Thatâs a mistake, a huge mistake. And from a taxpayerâs perspective itâs outrageous.â? We have a thirty-billion-dollar-a-year National Institutes of Health, he pointed out, which has been a remarkable powerhouse of discovery. But we have no billion-dollar National Institute of Health Care Delivery studying how best to incorporate those discoveries into daily practice.
Checklists are not the solution to every problem. A large portion of medicine deals with complex condition with large uncertainties and many disease interactions. Further, it may be more difficult for an insurance company to institute checklists than a hospital manager or someone further down the chain of command. Nevertheless, standardization in medicine should help to dramatically improve quality.
A recent paper by Franco et al. (2007) claims that increased poverty may improve health (see also NPR’s Marketplace report). How is this possible? Lower income reduces excess food as well as cigarette consumption. Further, poverty makes public transportation less affordable and individuals may substitute walking for taking the bus. The authors study Cuba’s experience between 1989 and 2005.
Cuba has been subjected to an economic embargo by the United States since the 1960s. After the loss of the Soviet Union as a trading partner in 1989, Cuba entered a prolonged economic crisis known as the ‘‘Special Period.’’ The crisis worsened continuously over the next 5 years, with economic output reaching a nadir in 1995 of about half the level in 1990.
The decreased economic activity lead to the following changes:
We see that the prevalence of many of risk factors declined during the “Special Period.” What were the affects on health?
So poverty is the answer? It turns out that all the news is not rosy:
Can these results be extended to other countries? Cuba is a communist country where health is provided publicly. The government can incur debt in order to provide medical care for its citizens. In the U.S., increased poverty will likely make medical care less affordable–for those not on Medicaid–and thus health outcomes may suffer. Further, long run economic decline will make the provision of even government-run high quality medical care unaffordable for a society. A final critique is that although the data presented in the paper is suggestive, correlation does not imply causation. One should always maintain a healthy skepticism regarding the conclusion of time-series correlation studies.While no economist would advocate for policy-makers to attempt to increase poverty, the Franco study may guide individuals into believing that “less is more,” at least when it comes to food intake and car usage.
The work of three researchers who developed the procedure of gene targeting have won the 2007 Nobel Prize in Medicine. The winners are:
Further coverage is available at:
What are the major differences between medicine and public health? What challenges do public health officials frequently ignore?
On Tuesday, I attended a seminar by Dr. Richard Schieber. Dr. Schieber was a practicing pediatrician, however for the last fifteen years he has worked as a medical epidemiologist for the CDC.
One of the major challenges facing the CDC is how to translate public health medical recommendations down to the clinical level. According to Schuster et al. (1998), little of public health care recommendations are preformed in the clinical setting. “Simple averages from a number of studies indicate that 50 percent of people received recommended preventive care; 70 percent, recommended acute care;… 60 percent, recommended chronic care.”
Why are these numbers so low? Well, public health officials have many recommendations. What if a physician is confronted with a patient who has diabetes and is also obese and a smoker. Will they be able to fulfill all the recommendations in the 15-20 minute time slot they have with their patient? Likely not.
A study by Yarnall et al. (2003) finds that “To fully satisfy the USPSTF [US Preventative Services Task Force] recommendations, 1773 hours of a physician’s annual time, or 7.4 hours per working day, is needed for the provision of preventive services. ”
Dr. Schrieber’s main point of the talk is that recommendations in and of themselves are not very useful. They must be “translated” so they are useful and feasible for both the medical provider and the patient. Further, public health programs should be measured by outcomes, not by processes.
Finally, some of the talk was spent on differences between public health and medicine. Some of these differences are listed below.
| Medicine | Public Health | |
| Scope | Individual | Populations |
| Boss | CEO or MD | Bureaucrat |
| Environment | Clinical | Desk Job |
| Salary | High | Not as high |
| Satisfiers | Positive patient outcome | Vague |
| Language | Very Technical (Greek) | English |
| Funding | Patient (via insurance or gov’t) | Government |
| Unbreakable rule | Primum non nocere | Help underserved |
| View of EBM | Dictatorial | Magically improve patient care |
| Basis of Decision | Patient history, physical, tests | Risk, QALYs, etc. |
Over the years, the CDC’s Advisory Committee on Immunization Practices (ACIP) has greatly increased the scope of the population for whom it recommends to receive an influenza vaccination. Currently, ACIP recommends that all children 6–59 months old, all health care workers, and all persons aged over 50–in addition to many other groups–should receive an annual flu shot.
However, MedPageToday reports (”…little mortality benefit…“) on an research by Simonsen et al. (2007) in The Lancet Infectious Disease journal. The Simonsen paper claims that influenza vaccination may be saving fewer lives than once thought among older adults.
The reason is that estimates of a 50% or greater reduction in all-cause mortality have emerged from cohort studies fraught with selection bias, asserted a review article in the October issue of The Lancet Infectious Diseases.
But the real effect with flu shots for those 65 and older during December through March could not have been any greater than 5% to 10%, said Lone Simonsen, Ph.D., of George Washington University here, and colleagues. That’s the flu-related mortality burden found in studies of excess all-cause mortality.
Aside from these cohort studies, the evidence is too weak to show any mortality benefit in older adults, who account for 90% of influenza deaths each year, Dr. Simonsen and colleagues added.
Why have other studies found that influenza vaccination had such a large effect of decreased mortality. Seriously ill people are less likely to get influenza shots; they have more immediate medical worries. These people are also more likely to die. Thus, we have a spurious, non-causal relationship between influenza vaccination and mortality due to selection bias.
The authors of the study wisely state the following:
“We must never again allow layers of poor research to mask substantial uncertainty about the effects of a public-health intervention and present a falsely optimistic view of policy,” they wrote. They called for placebo-controlled trials.
A new study by Tai-Seale, McGuire, and Zhang (HSR 2007) analyzes how primary care physicians allocate their time in a typical office visit. The authors use data from 392 videotaped office visits conducted in three settings: 1) a salaried group practice in an academic medical center (AMC), 2) a managed care group (MCG), and a fee-for-service inner city solo (ICS) practitioners with an Independent Practice Association contract. The authors found that the average length of a visit was 17.4 minutes but the median visit length was only 15.7 minutes.
| Total | AMC | MCG | ICS | |
| Median visit Length | 15.7 | 23.3 | 13.4 | 9.7 |
| Median time spent on Major topic | 5.3 | 6.7 | 4.8 | 3.2 |
| Median time spent on Minor topic | 1.1 | 1.4 | 0.9 | 0.7 |
The data above show that only about 5 minutes was spent on discussing the major issue facing the patient. For minor issues, the doctor and patient only spent one minute discussing the issue. We see that physicians at the AMC spent the most time with their patient while the physicians at the ICS spent the least. The authors find that “…while time spent by the patient and physician on a topic responds to many factors, time of the visit overall is much less malleable.”
The paper also notes that “[i]ncentives in prevailing physician payments favor procedure-based patient care over time-intensive evaluation and management care.” One could easily imagine a system with a flexible physician schedule. The patient could schedule a standard 15 minute appointment for the usual co-pay, or could pay the physician extra if they wanted a 20 minute, 30 minute or 1 hour appointment. This way, the physician would be reimbursed for their time. If the insurance company paid for these extra minutes, the physicians would have an incentive to exaggerate the number of minutes spent on each patient. Thus, a likely solution is for the insurance company to pay for the base appointment length (15 minutes) and anything over this the patient will have to pay for.
There is an interesting article from Sunday’s N. Y. Times magazine (”…what makes us healthy?“) about the problems in epidemiology of using non-randomized data to draw conclusions. For instance, there is much uncertainty as to whether hormone-replacement-therapy increases, decreases or has no effect on the probability a woman will have heart disease.
The article talks about a few important biases which occur when researchers analyze observational data:
There are also some interesting comments on Seth’s blog.
“GlaxoSmithKline Plc and Novartis AG, two of the world’s biggest vaccine makers, may have bet on the wrong technology in the race to develop a better flu shot.” This is how a recent article (”…Bet Wrong…“) by John Lauerman of Bloomberg News begins.
Helped by a $221 million grant from the U.S. government, Novartis is investing $600 million to build a cell-based flu-shot plant in North Carolina. Glaxo is receiving $275 million of U.S. government funding and plans to build a plant in Pennsylvania. Why may all these investments be going to waste?
Protein Science Corp. is close to gaining approval for using DNA to speed up the influenza vaccination development process.
“Protein Sciences inserts flu genes into cells from a corn- eating caterpillar. The cells are then induced to make proteins that when injected into humans trigger a protective immune response. In the technique Glaxo and Novartis are developing, mammal cells are used to generate whole flu viruses, which are then modified to make influenza vaccine….
The former Sanofi official [David] Fedson estimated that factories around the world using Protein Sciences’ recombinant DNA approach could make two doses of vaccine in three months for 3 billion people, almost half the world’s population. That’s 10 times as much as could be made in conventional egg-based flu- shot plants, he said.”
With worries of influenza pandemic increasing, this kind of innovation is a good sign for public health.
From The Economist (”Live fast, die young“):
“Rock stars are famous for excess, and some pay the price. A new study suggests that they are up to three times more likely to die young than the rest of the population, mainly because of drug and alcohol abuse.”
Click here to see a chart of the top causes of Rock Star mortality in the U.S. and Europe.
This is the question posed by a 2007 NBER working paper by Mary Beth Landrum, Kate A. Stewart, David M. Cutler. The authors use data from the National Long Term Care Survey (NLTCS) between 1994 and 1999. The data is panel in nature and has the benefit of combining Medicare administrative data with survey responses. The research examines individuals over 65 without disabilities in 1994 and see which patients became disabled and how this occurred.
Disability is defined by either:
Results
“Sixty-six percent of non-disabled respondents in 1994 survived and remained non-disabled to 1999, while 15.1% became disabled over the 5 year period and 18.9% died between survey waves. Six conditions made up about half of new disabilities (48%). These conditions were:
Only 17% of elderly respondents did not have one of these conditions and a majority (54%) had two or more. Of disabled respondents, 96% of newly disabled have at least one of these characteristics. Most disabled respondents had multiple conditions; two-thirds of newly disabled respondents had 3 or more conditions. The authors examine in more depth the interaction effects of having two or more conditions. For instance:
“…the interaction between diabetes and arthritis was positive, suggesting that these two conditions have synergistic effects such that having both conditions was more disabling than would be expected by the effects of each individual condition. In contrast, two interactions with dementia were negative (stroke*dementia and heart failure*dementia), suggesting that in the presence of a highly disabling condition like dementia, other conditions have effects that are dampened relative to what would be expected when the disease occurs in isolation.”
Hopefully, this data can be used to aid health service providers on how to better prevent and treat disabilities which occur in old age.
Can video games be used to learn how to best plan for infectious disease pandemics? Time reports on how a World of Warcraft pandemic can be used by epidemiologists:
[The] papers document the path of an unexpectedly virulent virtual disease called ‘Corrupted Blood,’ which swept through World of Warcraft’s online characters starting in September 2005. (The game’s administrators introduced the disease as a challenge for some high-level players; they didn’t expect it to break out of the caves and into the virtual world’s cities and towns.) The disease then ravaged the player population — despite administrators’ efforts to quarantine the infected — and gave World of Warcraft its first virtual-world pandemic.
In real life, epidemiologists have long used complex mathematical models to predict how an outbreak of, say, pandemic flu might spread around the world. The problem is that testing those models isn’t very easy, which makes it hard to judge whether the models are accurate. Scientists can’t just release pathogens into cities and see how many people die. So, instead, they base their models on past outbreaks, where information collection was imperfect, or on people’s stated (but hypothetical) beliefs about what they would do during a future outbreak. The resulting models can be remarkably sophisticated, but they “lack the variability and unexpected outcomes that arise … not by the nature of the disease, but by the nature of the hosts it infects,” according to Eric Lofgren and Nina Fefferman, authors of the Lancet Infectious Diseases paper. For example, they say, the failure of the World of Warcraft quarantine “could not have been accurately predicted by numerical methods alone, since it was driven by human decisions and behavioral choices.” In other words, no model will know whether or not people ignore infection-control rules in the real world.
Those interested in how diseases are spread may also find my 27 Jan 06 post interesting. It describes how flow of money may provide a good model of how diseases spread in modern society.
This questions seems very straightforward theoretically. If each appointment takes 20 minutes, then one should schedule patients at 8, 8:20 8:40, 9, etc.
The best laid schemes o’ Mice an’ Men,
Gang aft agley,
An’ lea’e us nought but grief an’ pain,
For promis’d joy!
In reality, however, this scheduling plan rarely functions smoothly. One may predict that patients arrive randomly around a mean of the scheduled appointment time. This would imply a Poisson process as long as arrivals occur one at a time, arrival patterns don’t change as the day progresses, and arrivals in disjoint portions of the day must behave in a statistically independent manner. A paper by Fontantesi et al. (2002) finds that this is not the case. Arrivals tend to be clustered. For instance, common bus schedules, traffic light timing, or parking space availability may lead to clumping of arrivals. Further, many people share similar lunch break hours and work schedules making it even more likely that arrivals will be clustered.
Arrival data collected in Fontanesi et al. (2002) show that the mean person is about 3-4 minutes early for an appointment. However, arrival distribution is heavily skewed to the right. While most arrivals fall between either 15 minutes early or 10 minutes late for an appointment, a non-trivial amount of people are extremely late for an appointment (more than 45 minutes).
What solutions do the authors propose?
Fontanesi J, Alexopoulos C, Goldsman D, DeGuire M, Kopald D, Holcomb K, Sawyer MH. (2002) “Non-punctual patients: planning for variability in appointment arrival times.” J Med Pract Manage., Jul-Aug;18(1):14-8.
The August 1st edition of JAMA has an interesting article which examines how exposure to war crimes affects individuals view about peaceful negotiations (”…War Crimes…“). The study takes place in the Acholi, Lango, and Teso subregions in northern Uganda. Since the late 1980s, many people in this area of Uganda experienced the bitter fighting between the Lord’s Resistance Army (LRA), and the Ugandan People’s Democratic Army.
The research paper interviews over 2500 adults in villages and camps for internally displaced person. The interviewers used various checklists to asses whether or not each individual suffered from depression and/or post-traumatic stress disorder (PTSD). Of the total sample, 45% suffered from depression and 74% were diagnosed with PTSD. Then the interviewers inquired as to the individuals’ views regarding peace.
“Respondents were asked how they defined peace. The answers were recoded at the analysis stage in 3 variables reflecting the range of responses: the absence of violence (yes or no); unity—eg, togetherness, living together (yes or no); and human/social development—eg, education, economic development (yes or no). Respondents were then asked how they believed peace could be achieved. The range of responses was recoded during the analysis stage in 2 variables: identified nonviolent mechanisms, such as peace talks and amnesties (yes or no); and identified violent mechanisms, such as killing enemy combatants and their leaders (yes or no).“
The study finds that “Respondents who met the PTSD symptom criteria were more likely to identify violence as a means to achieve peace (OR, 1.31; 95% CI, 1.05-1.65). Respondents who met the depression symptom criteria were less likely to identify nonviolence as a means to achieve peace (OR, 0.77; 95% CI, 0.65-0.93).”
The authors also divide respondents into 4 groups: 1) those with low exposure to violence, 2) those who had witnessed violence, 3) people directly threatened/injured and 4) a group who was abducted. Perhaps surprisingly, the groups did not differ significantly with regards to their views of peace–with the exception that group 4 was less likely to believe non-violent mechanisms could achieve peace. It seems to matter not just to which type of violence people were subjected, but also how they react mentally (with regards to depression and PTSD).
In summary, people who experience violent acts and suffer PTSD are more likely to view violence as a means to peace. This theory was voiced 50 years earlier in a less scientific, but more poetic manner by Dr. Martin Luther King, Jr. when he proclaimed: “Hate begets hate; violence begets violence.”
Accoding to a Canadian study, infants can tell the difference between two languages without hearing the spoken words. To read more on the study, see today’s Washington Post (”Babies Can Discern Languages…“).
Según una investigación canadiense, los infantes se pueden distinguir entre dos idiomas sin oÃr las palabras en voz alta.  Para más información acerca de la investigación, vea a La Razón (âLos bebés son capaces de distinguir un idioma a los cuatro mesesâ?).
Robin Hanson has spent the last week blogging on his Overcoming Bias website asking individuals to sign a petitions to redo the RAND Health Insurance Experiment (HIE). For those who do not know what the RAND HIE, Mr. Hanson has two posts (#1 and #2) describing the experiment. I agree with Mr. Hanson when he states “If you remember only one medical study, it should be the RAND health insurance experiment.”
But is a new RAND-style HIE needed? The Overcoming Bias site states that the purpose of a new HIE would be to “…try to make clear the aggregate health effects of variations in aggregate medical spending, variations induced by feasible regimes of quality control, including free patient choice induced by a varying aggregate price.” The proposal is for a study to randomly assign 10,000 people to different insurance types (or to no insurance) and follow them over 10 years. Mr. Hanson claims that this will cost 1/2 billion dollars over the 10 years. This seems reasonable, but possibly a little conservative—it is equivalent to purchasing $5000 worth of health insurance for 10,000 people each year for 10 years. Is it worth it?
As a researcher, I of course would love to be able to have access to new, high-quality data from a randomized controlled trial. On the other hand, I do not like the idea of spending 1/2 billion of taxpayer dollars (or private foundation money). One could be using these funds to feed the poor, provide medical care to the sick, or spend the money researching for new cures.
In the end, I think repeating the RAND HIE for the 21st century will be a worthwhile endeavor, and I have decided to to sign the petition. By conducting this RCT, we can see how much (if at all) health insurance improves health levels. A RAND-styled RCT could separately look at difference forms of care: preventative care, primary care and specialist care. According to a Health Affairs article by Catlin et al. (”National Health Spending in 2005“), government spending reached $902.7 billion in 1995. Of this, $342 billion was spent on Medicare and $313.1 billion was spent on Medicaid. With these astronomical figures at hand, one can see that 1/2 billion dollars over 10 years is only $50 million per year–less than 0.01% of public health care spending per year. If this cost figure is too high for some people, I think a sensible thing to do would be to reduce the sample size to between 7500-8000 people instead of the 10,000 proposed by Mr. Hanson. Nevertheless, Americans need to know what they are getting in exchange for their dollars used to finance Medicare, and Medicaid as well as their own purchases of private health insurance.
An interesting post by Arnold Kling (”Doctors, Pharmaceuticals, and Statisticians“) reports on a randomized clinical trial which demonstrated that on average, angioplasties have no incremental health benefits once the patient is placed on multiple medications such as beta-blockers, ACE inhibitors, statins and blood thinners.
Dr. Kling writes: “Doctors think that they add value by giving advice on issues such as angioplasty. But the advice of statisticians may be better. Doctors also think that drug companies earn too much money. But it may be the doctors who earn too much money. I predict a collision between doctors and statisticians somewhere down the road.“
According to the African Medical and Research Foundation (AMREF), “[m]alaria is the most important parasitic disease in the world. It kills 3,000 children every day and more than one million each year. The majority of these deaths occur among children under five years of age and pregnant women in sub-Saharan Africa.”
In the most recent edition of The Economist magazine, a piece titled “A shift of perspective” documents the research of Dr. Mauro Marrelli and Dr. Chaoyang Li of Johns Hopkins University. The researchers hope that breeding genetically-altered, malarial-resistant parasites and releasing them into wild will help to stem the tide of malarial infections. Drs. Marrelli and Li have found a gene, called SM1, which prevents malaria infections. SM1 is superior to genes which suppress but do not prevent infection since suppressing the infection “imposes a burden on the animal, since materials and energy have to be diverted to the boosted immune system. In practice, such insects are worse off than they would be without the gene. But because SM1 prevents infection rather than suppressing it, Dr Marrelli and Dr Li hoped that different rules might apply.” The mosquitoes with the SM1 gene may be superior (in a Darwinian sense) to those mosquitoes without it and thus more of these malaria-free mosquitoes will survive.
Problems do exist. It is possible that the SM1 gene is helpful only when one copy of the gene is present in a mosquito (and not two). Also, “the type of malarial parasite used in this and many other experiments is not actually one that causes human disease.” Still there is hope that scientific advances could one day rid the world of one of its most deadly parasitic diseases.
An interesting paper by Douglas Almond (2006) examines whether or not influenza infections of pregnant mothers can influence long-term outcomes of the in utero babies. Almond uses the 1918 “Spanish flu” pandemic in the United States as a source of exogenous variation to test the fetal origins hypothesis. The fetal origins hypothesis states that “certain chronic health conditions can be traced to the course of fetal development.”
The influenza outbreak of 1918 was brief (lasting only about 4 months), but deadly, killing more Americans than all combat deaths in the twentieth century. There were, however, 25 million individuals who contracted the deadly influenza strain and survived. Almond seeks to compare individuals born to pregnant mothers with the Spanish flu with a control group of individuals who were born from pregnant mothers without the virus. This is done two ways: using a dummy variable for being born in 1919 and using variation in one’s state of birth.
Using the 1919 dummy variable and comparing it to cohorts born between 1912 and 1922, Almond finds that the 1919 cohort had reduced educational attainment, increased rates of physical disability, lower income, lower socioeconomic status, and received more transfer payment than the surround cohorts.
In his secondary specification, the author examines 1960, 1970, and 1980 census data. Because census data include the state of birth for each person, Almond can trace use cross-state variation in infection rates from the 1919 Spanish flu. The author concludes that outcomes are worse for individuals born in 1919 in states with higher Spanish flu infection rates than for individuals born in 1919 in states with lower infection rates.
One may worry that the Spanish flu killed the less healthy or less ‘genetically endowed’ children in the 1918 cohort while the 1919 cohort survived because they were in utero. If this was the case, the 1918 cohort mean outcomes would be inflated and any differences may not be due to fetal development issues in the 1919 cohort, but the fact that the 1918 cohort may be more genetically endowed than usual. Almond finds that this is not the case, since the 1918 and 1920 birth cohorts have similar outcomes to the other birth cohorts in the 1912 to 1922 groups.
This paper gives robust evidence to support the fetal origins hypothesis. Prior epidemiological studies looked at famines, which tended to be more long-lasting and thus did not provide as much of a sharp, exogenous source of variation.
There are many myths in the popular press regarding what are the major determinants of increasing health care costs. Some pundits claim that illegal immigrants that do not purchase insurance are using emergency departments (ED) for primary care services, thus driving up the cost of health care for American citizens (see Science Daily’s interview with Jack Martin of the Federation for American Immigration Reform).
A study by Peter Cunningham (2006) in Health Affairs found this not to be the case. The author employs the 2003 Community Tracking Study (CTS) to study ED usage in 60 different U.S. sites. The findings are summarized in the chart below.
| High ED Usage | Low ED Usage |
| Citizens | Non-citizens |
| Black | White, Hispanic |
| Below Poverty Line | Above Poverty Line |
| Poor Health | Good Health |
| Live near ED | Live far from ED |
| Medicare/Medicaid insurance | Private insurance or uninsured |
| Long physician appt. wait times | Short physician appt. wait times |
| More outpatient visits/physicians | Less outpatient visits/physicians |
| More CHC revenue | Less CHC revenue |
Surprisingly, non-citizens have an ED visit rates which are 17.2% below those of an average American citizen. Among residents below the poverty line, non-citizens ED visit frequency is 30.3% below that of poor Americans. Also, while African Americans tend to use the ED more frequently, Hispanic ED usage is on par with that of whites.
It is also found that the uninsured have ED usage rates similar to individuals who hold private insurance. People with Medicare or Medicaid insurance use emergency room services much more frequently than the uninsured.
Some of the Cunningham’s conclusions include:
As the author suggests, “…reducing ED use defies simple solutions such as expanding insurance coverage or restricting access for undocumented immigrants.”
Cunningham (2006) “What Accounts For Differences In The Use Of Hospital Emergency Departments Across U.S. Communities?” Health Affairs. vol 25, no. 5, pp. w324-w336.
A push towards more accurate evaluation of health care quality has become a major public policy goal in recent years. The Department of Health and Human Services now has a website ranking hospital quality and California has a Healthcare Quality Report Card. One problem with simple quality measures such as mortality or morbidity rates are that physicians who have a patient base with a lower initial health level are penalized—according to the scoring—for treating the patients who need care the most. For this reason, health care rankings have moved from outcome-based rankings to process-based rankings. For instance, the California Report Card ranks insurance plans by such factors as whether or not individuals with diabetes had an eye exam and the percentage of pregnant women who had a check-up visit 21-56 days after delivery.
One puzzling finding Kahn et al. (2007) encounters in their data is that ambulatory care centers with better process of care scores have patient with a worse health outcomes. Using a simple ordinary least squares (OLS) procedure would lead to the erroneous conclusion that more care leads to worse health. Of course, individuals who are sicker are specifically the patients who need to undergo the most procedures.
To control for this problem, Kahn and co-authors use an instrument variables method. The independent variable—the process of care variables—is instrumented with a structure of care variable. The structure or care is defined in the data by a set of medical organization dummy variables. This IV estimation technique will provide unbiased, precise estimates if the medical organization dummies are correlated with the process of care, but the structure of care only affects health outcomes as mediated by the process of care. I believe that this is a reasonable assumption.
Using this methodology, the authors do in fact find that better processes lead to better health outcome. Moving a patient from a process of care score in the lower quartile to a process of care score in the highest quartile will lead to approximately a 10% increase in the PCS health score.
While the conclusion of this paper seems obvious—better process of care scores lead to better health outcome—the authors have done a yeoman’s job of proving this empirically. Further, the paper casts doubt upon the conclusions from other studies which use OLS regressions to compare health interventions and health outcomes.
In 2005, approximately 114 million visits were made by Americans to the hospital emergency departments. Of these, more than eighty percent concluded with a discharge and a recommendation for follow-up care. Receiving prompt and adequate post-ER care is imperative for the resolution of many illnesses and temporary disabilities. Is timely care available for these patients?
A study by Asplin, et al. (2005) and a subsequent paper by Neath and Carlin (2006) look at how easy it is to schedule an appointment after an ER visit. To collect the data, clinics were phoned by a graduate students posing as patients just released from a hospital emergency department. Callers had four (made-up) medical conditions: pneumonia, elevated blood pressure, vaginal bleeding in the first trimester, and symptoms of depression. The depression observations were excluded from the study because many primary care physicians do not feel qualified to treat depression.
In each call, the individual claimed to have either: 1) private insurance, 2) Medicaid insurance, 3) no insurance and could not pay, or 4) no insurance but would pay for the visit out-of-pocket. A call was deemed successful if an appointment was made within 7 days and the out-of-pocket payment for the appointment was $20 or less.
Results
Asplin, et al. preform a simple paired comparison in which the same clinics are compared where the only difference between the observations is the unit of insurance the phony patient had. Neath and Carlin directly incorporate other covariates – such as the medical condition, safety-net status of the clinic, city dummy variables, etc. The results are similar in both studies, but the table below gives Neath and Carlin’s findings.
| Clinic Type | Insurance Status | P(Success) |
| Non-Safety Net | Private | 68.4% |
| Non-Safety Net | Medicaid | 26.3% |
| Non-Safety Net | Uninsured | 14.6% |
| Safety Net | Private | 41.5% |
| Safety Net | Medicaid | 38.5% |
| Safety Net | Uninsured | 20.0% |
We can see that the “overall success probabilities in Asplin et al. were distressingly low.” One also notices that it is much easier to get an appointment if one has private insurance, but these differences are less severe at “safety net” clinics. Finally, the authors note that the majority of clinics made no attempt to determine the severity of the caller’s condition. Having trained staff answering the phone calls and preforming triage is costly, but is likely worth the cost for patients needing immediate assistance. Put more concisely, Asplin states: “Financial screening is trumping medical triage.”
With so much negative news on how ‘our health care system is failing,’ it is nice to see that “failure” may be an overstatement. According to the USA Today (”Cancer Deaths Drop“), the American Cancer Society reports that cancer rates have fallen for the second straight year.
This is great news. The National Center for Health Statistic’s 2004 Mortality Data show that cancer is the #2 killer in America. This table—cancer is listed as malignant neoplasms—shows that cancer causes half a million deaths per year, or 23.1% of all deaths in the U.S. Let us hope the trend of reduced cancer deaths continues.
A final analysis of the cost of flu vaccination is provided to us by Margaret Coleman, John Fontanesi and colleagues (2004). The authors examine the cost of the vaccination for different size practices in a scheduled visit and walk-in setting. Unlike most studies, this research team decided to applied overhead expenses to cost of vaccinations. While the marginal cost of overhead from a vaccination is zero, there is an incremental cost when the physician/firm decides to offer vaccinations. Vaccination storage, hazardous waste disposal, employee training and other fixed costs are incurred when the firm decides to vaccinate its patients. Thus, policymakers should take these costs into account if they wish to price services at average cost, but not marginal cost.
The costs of an influenza shot are as follows:
| Scheduled Visit | |||||
| Vaccine costs | Clinical Labor | Non-clinical labor | Overhead | Total | |
| Solo/Partner | $8.84 | $2.10 | $25.47 | $9.86 | $46.27 |
| Small | $8.51 | $6.05 | $10.32 | $9.57 | $34.45 |
| Medium | $8.19 | $2.00 | $8.10 | $9.48 | $27.77 |
| Large | $7.86 | $1.68 | $6.46 | $9.22 | $25.22 |
| Corporate | $7.54 | $1.35 | $1.55 | $9.14 | $19.58 |
| Walk-in Clinic | |||||
| Vaccine costs | Clinical Labor | Non-clinical labor | Overhead | Total | |
| Solo/Partner | $8.84 | $0.95 | $25.47 | $4.53 | $39.79 |
| Small | $8.51 | $2.75 | $10.32 | $4.42 | $26.00 |
| Medium | $8.19 | $0.91 | $8.10 | $4.32 | $21.52 |
| Large | $7.86 | $0.76 | $6.46 | $4.21 | $19.29 |
| Corporate | $7.54 | $0.61 | $1.55 | $4.17 | $13.87 |
The vaccination cost includes the actual cost of the vaccine, plus shipping, handling and storage, less any bulk discount. At the time of the publication of this paper, Medicare payment rates were only around $12-$17 per shot. This payment rate would have only covered average costs in the case of corporate firms operating walk-in clinics.
What is the best method to immunize individuals? Vaccinations are typically delivered via scheduled or walk-in visits. Mass vaccinations, however, may offer a more efficient means to vaccinate large populations. The mass vaccination locations can take place at schools, convention centers, fair grounds, churches, parking lots or other places (see Arkansas’ Mass Flu Vaccine locations as an example).
Authors John Fontanesi, Linda Hill and colleagues take a more in depth look at this issue in their 2006 paper. The researchers found that while only 59% of individuals who saw a physician for a well visit received a flu shot, 99% of those who went to a mass immunization clinic received the flu shot. The cost to administer the vaccines at the two sites were remarkably similar (see below).
Direct Unit Costs
| Routine Scheduled Visit | Mass Vaccination | |
| Physician ($61.43/hr) | $3.07 | $0.00 |
| Nurse ($21.93/hr) | $0.40 | $1.10 |
| Receptionist ($10.14/hr) | $0.24 | $0.68 |
| Vaccine cost | $8.73 | $8.73 |
| Syringe cost | $1.24 | $1.24 |
| Overhead | $5.04 | $6.30 |
| Total | $18.72 | $18.05 |
Thus, if we measure productivity as the cost for each person vaccinated, we see that mass clinics are about twice as productive. Regular office visits, however, are able to detect other health problems, whereas mass clinics only administer influenza vaccinations. Information gathering at mass vaccination clinics is minimal. The authors define efficiency as the quality of the visit divided by the cost of the visit. On this scale, the the study finds that office visits dominate mass vaccination clinics since they offer more patient education and better documentation of the encounter (see below). The authors conclude that “balancing critical resources…with the need to vaccinate as many at-risk patients…may be best accomplished by tightly integrating routine scheduled mass vaccination clinics with ambulatory care centers…”
Productivity (The lower score, the greater the productivity)
| Routine Scheduled Visit | Mass Vaccination | |
| Mean Labor Cost | $3.72 | $2.45 |
| Mean Overhead Cost | $5.04 | $9.90 |
| Mean Materials cost | $9.98 | $9.98 |
| Mean Patients seen/hr | 2.87 | 8.9 |
| Production Value | 6.58 | 2.50 |
Efficiency (The higher the score, the greater the efficiency)
| Routine Scheduled Visit | Mass Vaccination | |
| Nbr who could be vaccinated/hr | 2.87 | 8.9 |
| Probability of being vaccinated | 59% | 99% |
| Probability of health history reviewed | 59% | 0% |
| Probability VIS given | 54% | 71% |
| Probability vaccination documented | 70% | 45% |
| Efficiency Value | 0.020 | 0.001 |
A second paper by Fontanesi and colleagues published in 2004 uses critical path analysis to analyze the rate of influenza immunization at 1) scheduled visits, 2) walk-in visits, 3) scheduled shot only visits, and 4) walk-in shot only. They find vaccination rates similar to the 2006 paper listed above. Immunization rates are 58%, 45%, 100%, and 98% respectively. The main indicators which contributed to a vaccination occurring were 1) the patient and provider discuss immunization, 2) the ratio of the number of providers to the number of registration staff, 3) the ratio of the number of providers to the number of exam rooms, and 4) the patient was asked about the examination in the exam room or before the exam. The authors wisely conclude the following:
“Critical path analysis of vaccination activities occurring during routine scheduled health encounters suggests that this is a complex task and should not simply be seen as an incremental activity added to the general health encounter. Critical path analysis also suggests that provider-patient discussion is more productively viewed as the culmination rather than the beginning of the vaccination process.”
Yesterday we examined the cost effectiveness of the influenza vaccine in children. Today I will look at two papers which focus on the same question using children as the sub-population of interest. Estimated influenza infection rate among young healthy children is between 35% and 50% each year.
Cohen and Nettleman (2000) look at preschool children and create a cost-benefit scenario under a flexible and fixed (i.e.: 8am – 5pm) schedules. The net cost savings for the vaccination is calculated as below.
| No Vac | Flexible | Fixed | |
| Direct Cost | 14.87 | 14.53 | 14.53 |
| Indirect Cost | 29.16 | 8.23 | 28.3 |
| Total Cost | 44.03 | 22.76 | 42.83 |
| Net Cost | -21.27 | -1.20 | |
The authors find that direct costs of administering the vaccination can be as little as $4 in an HMO, and up to $10 in an outpatient stand-alone practice. Cost savings from the vaccination results from fewer pediatric office visits (estimated to cost $51/visit) and fewer ER visits (estimated at $124.42/visit) as well as fewer antibiotics used and fewer hospitalizations. Indirect costs included the parental time spent in the doctor’s office with their child as well as time absent from work from having to take care of a sick child.
This analysis shows that flexible vaccination is far superior to fixed schedule vaccination. The problem with this is that it is likely that health care providers would need to be compensated more to work a flexible schedule which may decrease the cost savings. Also, vaccination costs do not count physician rent, utilities, staffing, chart maintenance, etc. which go into the cost of a physician visit.
A second, more reliable study is preformed by Luce, et al. (2001). This study employs data from a multicenter, prospective, randomized, double-blind, placebo controlled efficiency trial during two flu seasons (fall 1996 through spring 1998). The vaccine studied was the live, attenuated, trivalent, intranasal influenza vaccine (not the more commonly used intramuscular shot). It was found that vaccinated children have 1.2 fewer influenza related illness (ILI) fever days/child than unvaccinated children. The costs of the vaccine are analyzed are as follows:
| COST CATEGORY |
| Direct Costs |
| Resources to administer vaccination |
| Resources to treat vaccine-associated adverse affects |
| Resources to treat ILI |
| Direct non-medical costs |
| Caregivers’ travel for vaccination |
| Caregivers’ travel to health care visits from vaccine-related adverse affects |
| Caregivers’ travel to health care visits for treatment of ILI |
| Lost Productivity Costs |
| Caregivers’ missed usual activity to receive vaccine |
| Caregivers’ missed usual activity as a result of vaccine associated adverse affects |
| Caregivers’ missed usual activity as’ a result of children’s ILI |
For the direct costs, the authors take into account the vaccine cost, time to administer the vaccine, and medication use from adverse affects of the vaccine. To measure the health care resources used to treat ILI, the authors look at expected costs from hospitalization, health care provider visits (ER or physician), laboratory and diagnostic tests and procedures, antibiotics and other prescription or over-the-counter medications.
Direct non-medical costs and lost productivity costs are measure using transportation costs (e.g.: vehicle mileage and parking costs) and caregivers’ time costs respectively. Most of the cost data is culled from other papers or NAMCS and NHAMCS data. The authors find that from society’s point of view, a $28 intranasal vaccination was worth the cost. From an individual’s point of view, an intranasal vaccination costing less than $5 was worth the cost. The individual’s break even value is lower than society’s since the individual does not take into account the externality that if they are vaccinated, it is less likely that someone else becomes infected with influenza.
How does one measure the cost effectiveness of flu vaccination? Today we will examine a paper which attempts to do just that.
Nichol, et al. (2003) preforms a cost-benefit analysis of live attenuated influenza vaccination (LAIV) for healthy, working adults. This type of vaccination is administered through a nasal spray than through an intramuscular shot. The authors attempted to answer this question using a multi-center, randomized, double blind placebo controlled trial. They found that adult influenza vaccinations decreased the number of work days missed by 18%, decreased the number of less effective work days by 18% and decreased the number of provider visits by 13%. This amounts to a large benefit to society, but one must also estimate the costs of vaccination in order to have an accurate cost-benefit analysis. Nichol and colleagues estimate the costs as follows:
| COST CATEGORY | EQUATION |
| Costs of vaccination | |
| Direct costs of vaccination | Costs of vaccine and administration |
| Indirect costs of vaccination | P(missing work for Vac.) x (hrs of work missed for Vac.) x (wage) |
| Direct costs of side effects due to vaccination | P(side effects)x (provider visit cost) |
| Indirect costs of side effects due to vaccination | P(side effect lead to work loss)x(wage)x(8 h/day) |
| Costs averted due to vaccination | |
| Direct costs of medical care averted due to vaccination | P(provider visit from illness among NV) x (1 − RR_V) x (cost of provider visit) |
| Indirect costs of work loss averted due to vaccination | P(work loss due to illness among NV)x(1 − RR_V)x(wage)x(8h/day) |
| Indirect costs of reduced work effectiveness | P(reduced work effectiveness_NV) x (1- RR_V) x (wage) x (8h/day) |
Provider visit costs is estimated separately for influenza-like illness (ILI) and upper respiratory tract illness. Costs include not only physician time, but also the average number of tests and procedures ordered and medications prescribed (e.g.: antibiotics, Rx and over the counter meds) and are estimated using NAMCS data. The authors find that the break even cost of flu vaccinations for healthy adults is $43.07. Monte Carlo simulation is then used to prove their results more robust.
The study does have some problems. The authors do not include overhead costs in their estimate of physician visits. Secondly, no justification is found for the distributional assumptions of the Monte Carlo simulations. Most distributions are triangular between the min and the max.
According to Thompson, et al. (2003), approximately 51,000 people per year died annually due to influenza related diseases between 1990 and 1999. Mortality rates are appreciably higher for those over 65 years of age.
In order to reduce mortality and morbidity from influenza in the U.S., the CDC’s Advisory Committee on Immunization Practices (ACIP) released its “Prevention and Control of Influenza” guide this summer. The report documents important information regarding influenza and gives recommendations to patients and providers regarding when/for whom/in what manner influenza vaccinations should be administered.
According to this paper, influenza is divided into two types: influenza A and influenza B. Influenza A is further divided into two subgroups: hemagglutinin and neuraminidase. Influenza represents a unique family of viruses since antibodies developed against one strain of influenza offer little or no protection from other strains. There are also two types of vaccines. The live attenuated influenza vaccine (LAIV) uses live but weakened influenza virus. LAIV is administered through a nasal spray. The second, more traditional vaccine is the inactivated vaccine which is administered through an intramuscular injection and uses killed viruses. While the LAIV is often less unpleasant for the patient, it is usually more costly than the inactivated vaccine.
Hospitalization Rates
A variety of studies have found the following hospitalization rates which occur as a result of influenza. Below are a summary of a few studies which analyze the American population.
| Years | Population | Age | Hosp/100k – High Risk | Hosp/100k – Healthy |
| ‘73-’93 | TN Medicaid | 0-11 mos | 1900 | 496-1038 |
| 1-2 yrs | 800 | 186 | ||
| 3-4 yrs | 320 | 86 | ||
| 5-14 yrs | 92 | 41 | ||
| ‘92-’97 | 2 HMOs | 0-23 mos | 144-187 | |
| 2-4 yrs | 0-25 | |||
| 5-17 yrs | 8-12 | |||
| ‘68-’69 | HMO | 15-44 yrs | 56-110 | 23-25 |
| ‘70-’71 | HMO | 45-64 yrs | 392-635 | 13-23 |
| 72-’73 | HMO | >64 yrs | 399-518 | |
| ‘69-’95 | Nat’l Hosp Data | <65 yrs | 20-42 | |
| ‘69-’95 | Nat’l Hosp Data | >64 yrs | 125-228 | |
| ‘79-’01 | Nat’l Hosp Data | All | 88 | |
New Recommendations
The influenza vaccine is now recommended for a wide range of people (see “Persons for whom annual vaccination is recommended” below). Three important changes in ACIP’s vaccination policies for 2006 are:
Children under 6 years and parents of children under 6 should receive an annual flu vaccine.
Children under 9 years of age who have never been vaccinated should receive 2 doses of the vaccine.
Continued offering of the influenza vaccine even after influenza activity has been documented in a community.
Persons for whom annual vaccination is recommended
Children aged 6–59 months;
Women who will be pregnant during the influenza season;
Persons aged >50 years;
Children and adolescents (aged 6 months–18 years) who are receiving long-term aspirin therapy and, therefore, might be at risk for experiencing Reye syndrome after
influenza infection;
Adults and children who have chronic disorders of the pulmonary or cardiovascular systems, including asthma (hypertension is not considered a high-risk condition);
Adults and children who have required regular medical follow-up or hospitalization during the preceding year because of chronic metabolic diseases (including diabetes
mellitus), renal dysfunction, hemoglobinopathies, or immunodeficiency (including immunodeficiency caused by medications or by human immunodeficiency virus);
Adults and children who have any condition (e.g., cognitive dysfunction, spinal cord injuries, seizure disorders, or other neuromuscular disorders) that can compromise respiratory function or the handling of respiratory secretions, or that can increase the risk for aspiration;
Residents of nursing homes and other chronic-care facilities that house persons of any age who have chronic medical conditions;
The important changes
Thompson,; Shay; Weintraub; (2003) “Mortality associated with influenza and respiratory syncytial virus in the United States.” JAMA, vol 289, pp. 179-186.
Which treatment is more effective: placebo one or placebo two? This is the test researchers Ted Kaptchuk and colleagues investigated in their study published in the British Medical Journal and summarized in an April 2006 issue of Discover (”Placebo vs. Placebo“). Kaptchuk looked at 266 volunteers with chronic arm pain and randomly assigned each of them to either receive a sugar pill or be subjected to pretend acupuncture.
Both groups experienced side effects to the placebos:
“25 percent of the acupuncture group experienced side effects from the nonexistent needle pricks, including 19 people who felt pain and 4 whose skin became red or swollen. 31 percent of the pill group experienced side effects from the make-believe drug, including dizziness, restlessness, rashes, headaches, nausea, and 4 cases of nightmares. Dry mouth and fatigue were the most common side effects, and 3 subjects withdrew from the study after reducing the dosage failed to control their symptoms. The reported side effects exactly matched those described by the doctors at the beginning of the study.”
After ten weeks, individuals receiving the fake acupuncture reported a larger reduction in pain than those taking the sugar pill. Kaptchuk has a logical explanation for this.
“Kaptchuk says that the rituals of medicine explain the difference: Performing acupuncture is more elaborate than prescribing medicine. Other rituals that may make patients feel better include ‘white coats, and stethoscopes that you don’t necessarily use, pictures on the wall, the way you reassure a patient, and the secretaries that sign you in.’ Careful manipulation of such rituals could make all types of treatment more effective, Kaptchuk suggests.“
What is the most efficient procedure to immunize large groups of people? This is an interesting question, especially considering the potential need to distribute vast amounts of medicine in the case of a terrorist attack. Since the CDC’s Vaccinations for Children (VFC) program recommended flu shots for that all children age six months and to 5 years, and many children between ages 5 and 18, we can see how best to vaccinate a large group of people in this setting. Influenza vaccinations are only effective for one year, so there will be a recurring demand each fall.
On Saturday I observed a mass immunization at a suburban San Diego health care facility. Without question, these mass immunizations were significantly more efficient than flu shots distributed via a walk-in clinic setting or a typical well-child visit. The entire procedure (i.e.: taking the patient’s temperature, performing basic paperwork and having the nurse administer the shot) only took about 10-15 minutes. No physicians were needed to supervise the process and thus this system was much less expensive. Well-child visits typically take at least 45 minutes when you take into account the time spent in the waiting room, time spent with a nurse in the pre-exam room, and the time spent with the physician.
Pharmacies also offer a cost-effective setting for immunizations. Most pharmacies are conveniently located for consumers and the cost to administer the shots in this setting are lower than the costs in a physician’s office. Further, firms such as Long’s, CVS, and others believe that flu shots are an effective means to create foot traffic and customer loyalty.
There are two treatment options for patients with breast cancer. The first is a breast conserving surgery (BCS) which removes the cancerous lump (lumpectomy) followed by irradiation treatment. The second option is a mastectomy which removes the entire breast. Lecia Apantaku claims in the American Family Physician journal in 2002, that “survival rates following breast conservation surgery plus radiation are equivalent to those following mastectomy.”
A 2003 Health Services Research paper by Hadley, et al. hypothesizes that differences in Medicare fees for the two procedures affects the probability that one or the other procedure will be used by the physician. To prove this point, the authors use 1994 Medicare claims data for a sample of 1,787 Medicare patients who were treated for early stage breast cancer. The authors compare Medicare fee rates in various geographic areas. The prediction is that in areas with relatively higher physician compensation for BCS, there will be more BCS procedures performed. In areas with relatively higher surgery fees for mastectomies, the opposite will hold. The authors also take into account: input costs using HCFA values of the Geographic Adjustment Factor (GAF), physician year of graduation to control for surgical preferences of a cohort, whether the patient had supplemental insurance, as well as various demographic variables. A mutinomial logit regression framework was use in which the dependent variables were: 1) BCS only, 2) BCS with radiation, and 3) mastectomy.
Results
The authors find that Medicare fees were significant factors in the choice between mastectomy and breast conserving surgery with radiation. A ten percent increase in the BCS with radiation fee (i.e.: about $30) increased the relative odds of BCS with radiation relative to a mastectomy to 1.34 (p-value 0.02). A 10% decrease in the mastectomy fee increased the relative odds of BCS with radiation relative to a mastectomy to 1.84 (p-value <0.01). The affect of fees for 'BCS only' did not lead to a statistically significantly impact on the probability of having a BCS only procedure, but this may be due to the fact that BCS only procedures are relatively infrequent.
Conclusion
The authors claim that this evidence “is consistent with the hypothesis that physicians are responsive to financial incentives when the alternative procedures have clinically equivalent outcomes, and patient clinical condition does not dominate the treatment choice” [my emphasis]. The authors wisely note there could be a reverse causality here. Physicians may tend to charge more for the procedure which is performed most often in the region. Other problems are that the data sample is small, the data are over ten years old, and the authors do not model patient preference for one procedure over the other.
Hadley; Mandelblatt; Mitchell; Weeks; Guadagnoli; Hwang; (2003) “Medicare breast surgery fees and treatment received by older women with localized breast cancer” Health Services Research vol 28, no. 2, pp. 553-573.
People love their doctors. Study after study has shown this to be true. Typically 75%-95% of patients surveyed claim that they completely or mostly trust their physician. Does this finding hold even when doctors are compensated through a capitation method and have a financial incentive to withhold care? A study by Kao, et al. (1998) shows that even capitation patients largely trust their doctor. The authors find 94% of FFS indemnity patients, 85% of managed FFS patients, 83% of salary patient patients and 77% of capitated patients trust their doctor.
A study by Hall et al. used a randomized trial where half of HMO patients were explicitly informed that their doctors were compensated on a capitation basis and had a financial incentive to withhold services. These disclosures were found to have no negative effects on the trust of either physicians or insurers. The authors conclude that “once patients have formed opinions of their physicians based on actual experience with them, information about payment methods has little or no effect on their opinions, possibly because of the cognitive dissonance that otherwise would arise.” It would be interesting to examine whether a doctor’s actual proficiency or the doctor’s personality mattered more to the patient in terms of their level of trust.
Hall; Dugan; Balkrishnan; Bradley; (2002) “How disclosing HMO physician incentives affects trust” Health Affairs, March/April, pp. 197-206.
Kao; Green; Zaslavsky; Koplan; Cleary; (1998) “The relationship between method of physician payment and patient trust” JAMA, Vol 280, No. 19, pp. 1708-1714.
Are you feeling ill or experiencing worrying symptoms? Interested in looking for information regarding a disease? If so, you can find answers to these and other medical questions at Wrong Diagnosis. This site uses reliable sources to create it medical database. These sources include:
Can Pay for Performance (P4P) improve care and slow spending growth in the U.S.? Joe Paduda is doubtful. So is John Wennberg of the Dartmouth Medical School. In his report (”Variation…“) for The Commonwealth Fund, Wennberg says that P4P initiatives can be very effective in creating incentives for physicians to operate under best practices. For instance, practice guidelines are for diabetics to have an eye examination at least every two years. We could compensate physicians more whose diabetic patients made visits on average every two years.
The problem Wennberg brings up is that the money spent on procedures where there are clearly defined best practices is small compared to spending in areas where there is no established ‘optimal’ protocol. Wennberg divides medical procedures into 3 groups.
“Preference-sensitive care typically involves significant tradeoffs that affect the patientâs quality or length of life. The surgical options for treating early stage breast cancer, for example, usually include mastectomy (complete removal of the breast) or lumpectomy (a local excision of the tumor), often called âbreast-sparing surgery.â? The consequences for women who choose mastectomy include the loss of the breast and, for some, the use of a prosthesis or the undergoing of reconstructive surgery. For women who choose breastsparing surgery, consequences can include radiation or chemotherapy, or both, and living with the risk of local recurrence, which would require further surgery.”
Wennberg estimates 50% of all medical spending is for supply sensitive care. This tends to suggest that P4P should be promoted, but only on a narrow scale and only for procedures with clear best practice guidelines. P4P may be a step forward but it is not a cure all for the modern medicine’s problems.
Good news for coffee-aholics like myself. The Seattle Times reports (”Coffee’s Health Conundrums“) that coffee may have health benefits including a reduced risk of type 2 diabetes. Researchers at the Pauling Institute concluded that “there is little evidence of health risk and some evidence of health benefits” for up to four cups a day.
Ten days ago, MedPageToday ran an article (”Hefty Bank Account…“) which claimed that people who have more money are healthier. Using the 2000 Census American Community Survey, the study finds that “a 55-year-old man making about $49,500 per year is 44% more likely to have a functional disability than his neighbor making $57,800 a year.” This finding is entirely believable, yet does not really tell us what we want to know: what is causing good or bad health?
Does more money enable an individual to purchase better health care and thus increase the probability of desireable health outcomes? Is increased earning power an indicator for socio-economic status, better education, etc. which may lead people to adopt healthier life-styles? It is also possible to find reverse causality here where people who are ill are less productive at work and thus earn less. The authors of the study even admit that since their data is cross-section, their study is not able to determine the direction of causality. If the investigators used panel data, their study could track the individuals over time and see what are the major causes of their illnesses.
Speaking about which services hospitals and doctors should provide at the end of a patient’s life is always a sensitive subject. Family members become emotional and desire any procedure which has a non-negative chance of prolonging the patients life, regardless of the cost. This is what insurance is for, to safeguard families from calamitous medical conditions. Having the end of life procedures covered by insurance, however, means that many unnecessary services are being provided. The Associated Press gives an example in the case of cancer patients (”Cancer patients, doctors don’t know when to give up“).Â
Overly aggressive treatment gives false hope and puts people through grueling and costly ordeals when there is no chance of a cure, cancer specialists said.
     âThere is a time to stop,â? said Dr. Craig Earle of the Dana-Farber Cancer Institute and Harvard Medical School. âItâs sometimes easier to just keep giving chemotherapy than to have a frank discussion about hospice and palliative care.â?
Â
After examining over 200 thousand Medicare medical records of cancer patients who died in the 1990s, Dr. Earle found ICU visits in the last month of life climbed from 8% of patients in 1993 to 11% of patients in 1999, although the number of patients visiting hospice care facilities in the last three days of life increased 3 percentage points as well. Even the president of the National Coalition for Cancer Survivorship said:
âI see, in cancer care, so much treatment being used in the last three months of somebodyâs life that doesnât really help.”
Mandating what care patients should have is not desirable. Doctors should take into account the cost of procedures when they preform them; however as the system is currently set up, physicians have no incentive to reduce costs (except maybe in HMOs).  This is a problem for which I currently have no good solutions. Any suggestions?Â
According the CIA World Factbook, life expectancy at birth in the U.S. has reached nearly 78 years old. In Singapore, Hong Kong and Japan, life expectancy is over 81. One wonders, however, how much of the increase in longevity in recent years is due to rising incomes, better diet and exercise, and how much is due to medical advances.
Some experts would say that the latter has contributed little. Dr. David Eddy claims in a Businessweek article (”Medical Guesswork“) that only 20-25% of medical procedures have proven to be effective. In Dr. Eddy’s words “The problem is that we don’t know what we are doing.” The article also gives the following anecdote:
“With a groundbreaking computer simulation, Eddy showed that the conventional approach to treating diabetes did little to prevent the heart attacks and strokes that are complications of the disease. In contrast, a simple regimen of aspirin and generic drugs to lower blood pressure and cholesterol sent the rate of such incidents plunging…
The message got through. Three years later, Kaiser is in the midst of a major initiative to change the treatment of the diabetics in its care. “
Dr. Eddy isn’t the only expert to voice concerns that much medical care may be ineffective. Discover magazine (”Dialogue“) has an interview with Dr. Nortin Hadler. Dr. Hadler claims that bypass surgery is never appropriate. His explanation of why the surgeries are still conducted is the following:
“Money. Interventional cardiology is what supports almost every hospital in America—it’s an enormous part of our gross domestic product. Every year in this country we do about half a million bypass grafts and 650,000 coronary angioplasties, with the mean cost of the procedures ranging from $28,000 to $60,000.”
Dr. Hadler’s main determinate of life expectancy is also interesting:
“With regard to socioeconomic status, the central question relates to relative wealth—in other words, the smaller the income gap in a given area, the better the longevity. Where the income gap is larger, the poor die sooner. These are powerful associations. The answer does not lie in modern medicine but in modern society.”
This association may (or not) be true for developed countries. Anecdotally, it does hold in countries like Cuba where individuals are poor ($3,300 GDP/capita), but inequality is low. Life expectancy in Cuba is only 0.44 years behind that of the U.S. I believe, however, that in general increasing GDP/capita generally increases life expectancy in developing countries even if inequality increases.
Further, I posit that wealth is highly correlated with good health, but it is difficult to tell if these relationships are causal. Does good health allow a worker to make more money during their lifetime, or does more wealth allow a worker to purchase better health care? The direction of causality is indeterminate.
According to the PhRMA (Pharmaceutical Research and Manufacturers of America) U.S. drug companies spent $39.4 billion on research and development in 2005. Much of this money goes towards the clinical trials necessary for FDA approval. But how much does it cost to bring a drug to market?
In order to bring a drug to market, a firm must go through a variety of phases for FDA approval. There are pre-clinical trials on animals. Next, in phase I, a small number of healthy volunteers are tested in order to establish safe doses and to gather information on the compound. In phase II, 100-300 individuals with the disease are selected in order to determine safety and efficacy. Phase III repeats phase II, but instead uses a sample of 1000 to 3000 individuals and examines the long run health effects of the drug as well.
DiMasi, Hansen and Grabowski (2003) estimate the cost of bringing a drug from phase I to market. Their data come from the Tufts Center for the Study of Drug Development (CSDD). The firms included in their survey represent 42% of all pharmaceutical R&D expenditures in the U.S. The authors found that the time from the start of clinical testing to marketing approval was approximately 90.3 months. This figure is in addition to any development time which occurs before phase I clinical trials. The authors take into account the cost of money used to finance the R&D using a 9% real cost of capital estimate. The final estimate is that it costs–including the expense of failed drugs–$802 million to take a drug from phase I trials to approval. Over 50% of this figure is the cost of capital needed to finance the R&D over such a long period.
Some observers would say that reducing FDA restrictions would reduce the price of drugs consumers face. I do not believe this to be the case. After the R&D is spent, firms price their drug to maximize profits subject to consumer demand. Reducing R&D costs will reduce the sunk costs, but not the marginal costs for pharmaceutical producers. What reducing FDA restrictions will do is increase a firm’s incentive to invest in drug development, because the revenue threshold to make an adequate return on capital will be reduced with a lower cost of gaining approval.
DiMasi, Hansen, Grabowski (2003) “The price of innovation: new estimates of drug development costs,” Journal of Health Economics; Vol 22, pp. 151-185.
The concept of the Physician Assistant gained its inspiration from 17th century Europe where feldshers were used in the 17th century Russian Army. In the 1960s, China employed over 1.3 million “barefoot doctors” to improve delivery of health care, especially in rural areas. Not until the mid 1960s did the U.S. begin to use Physician Assistants to deliever medical care due to a shortage of primary care doctors.
In the United States, Physician Asssitants (PAs) must be associated with a physician and must practice in an interdependent role. The partner physician, however, does not need to be physically present during a PA examination of a patient. PAs routinely deal with uncomplicated sprains, strains, hypertension, bronchitis, depression, allergies, asthma, gynecological problems, family planning and trauma. Approximately 55% of all physician assistants practice in primary care.
In order to become a Physician Assistant, the average PA spends 25 months studying an intensive core curriculum. In 2001, there were 130 training programs in universities, medical schools, colleges, and the armed forces. PAs learn the broad topics related to primary care and rotate through the major specialties. Nurse practitioners, on the other hand, traditionally are trained in one specialty (pediatrics, women’s health, etc.).
The following are some summary data for Physician Assistant which comes from the American Academy of Physician Assistants 2005 Census.
Number of Physician Assistants by Disorder in 2005
BY PRIMARY EMPLOYER
| Single-specialty physician group | 30.6% |
| Other hospital | 14.9% |
| Solo physician practice | 13.5% |
| Multi-specialty physician group | 12.3% |
| University hospital | 7.5% |
| Community health center | 6.1% |
| Self-employed | 3.1% |
| HMO | 2.3% |
| Other | 9.7% |
BY GENERAL SPECIALTY PRACTICED
| Family medicine | 28.4% |
| Surgical subspecialties | 21.9% |
| Other | 10.5% |
| Internal medicine subspecialties | 10.3% |
| Emergency medicine | 9.7% |
| General internal medicine | 7.6% |
| General surgery | 2.8% |
| General pediatrics | 2.5% |
| Obstetrics & gynecology | 2.4% |
| Occupational medicine | 2.3% |
| Pediatric subspecialties | 1.5% |
ANNUAL INCOME (Full-time workers only)
| Mean | $81,129 |
| 10th percentile | $60,184 |
| 25th percentile | $67,128 |
| Median | $77,402 |
| 75th percentile | $90,402 |
| 90th percentile | $106,705 |
AAPA 2005 Census
Mittman, Cawley, Fenn; (2002) “Physician Assistants in the United States,”British Medical Journal, Vol 325, 31 August 2002.
Nearly every day one reads about a revolutionary new pharmaceutical or medical procedure. Years later, however, we often learn that this ‘breakthrough’ was only a marginal improvement, had serious side effects or simply did not work. Ellen Goodman discusses in “Health ‘breakthrough’ anxieties” how a new breast cancer drug (raloxifene) seemed to offer a better choice to post-menopausal women at risk of breast cancer than the previous alternative (tamoxifen). Although raloxifene did seem to outperform tamoxifen, Cynthia Pearson of the National Women’s Health Network figures that only 30 out of the 10,000 who too raloxifene for up to five years actually benefited. The oncologist Jerome Groopman puts forth an appropriate newspaper headline:
“There’s a Small Difference in a Larger Group of Women That Has Side Effects and It’s Not Clear What’s Best for Any Individual“
The point is to be skeptical of what you read in the newspapers. If you believe you need some sort of medication, it is your responsibility to yourself to the necessary research to find out which drugs (or medical procedures, etc.) are best for your particular situation.
Most people intuitively believe that having more nurses on staff at a hospital improves health outcomes. After reading Money Magazine’s report that an average RN earns approximately $70,000 per year, relying on ‘intuition’ may not be the most appropriate manner to judge a nurse’s cost effectiveness. Do health outcomes really improve to justify this cost?
Needleman, Buerhaus, Mattke, Stewart, and Zelevinsky (2002) provide convincing evidence that nurses do improve health outcomes in hospitals. The study examines 799 hospitals in eleven states and tests to see how variation in nurse staffing or nurses hours worked changes health care outcomes. The first stage of their analysis runs a logistic regression. The health outcome for each patient was regressed on patient diagnosis related group (DRG), age, sex, primary insurer, state of residence, a dummy for emergency admission and the presence of any chronic diseases. These factors were added up and each hospital was assigned a risk factor. The second stage uses an ordinary least squares (OLS) regression to calculate the difference between the expected health outcome and the actual outcome using hospital dummies, nurse staffing and hours, the number of beds, etc.
Below is a table of their results. All coefficients less then 1.00 indicate that health outcomes improved. Outcomes statistically different from 1 at the 5% level receive a star (*).
| Proportion of RN hours | No. of RN hours/patient day | |
| Length of Stay | -1.12* | -0.09* |
| Urinary Tract Infection | 0.48* | 0.99* |
| Upper gastrointestinal bleeding | 0.66* | 0.98* |
| Hospital-acquired pneumonia | 0.59* | 0.99* |
| Cardiac Arrest | 0.46* | 0.98 |
| Failure to Rescue | 0.81* | 1.00 |
| In-Hospital Death | 0.90 | 1.00 |
Since outcomes improve in all six of the seven categories under the first model and in four of seven categories under the second model, it seems that nurses do have a positive effect on health.
I find two possible problems with the study:
The first stage logistic regression may have omitted variables. For instance, if rich people have better health outcomes—due to lifestyle choices—and are able to afford hospitals with more nurses, we would find a spurious correlation between health and nurse staffing, since nurse staffing level is simply a proxy for inherently healthier patients.
Despite these two problems, the evidence does seem convincing. As standard economic theory would predict, increasing inputs (nurses) will lead to increased outputs (health) ceteris paribus. Future research needs to determine more precisely what is a nurse’s cost benefit ratio in order for hospitals to ascertain the appropriate RN staffing.
Needleman, J; Buerhaus, P; Mattke, S; Stewart, M; Zelevinsky, K (2002); “Nurse Staffing Levels and the Quality of Care in Hospitals,” New England Journal of Medicine, Vol 346 (22).
The problems with the Medical Malpractice system in the US have been well-documented. President Bush has presented proposals to cap punitive damages in malpractice litigation. Other others have decried the fact that despite a large number of negligence cases each year, very few patients bring suit to court. Below are two studies which should give the reader a more informed perception of how malpractice law functions in the United States today.
This first study is by Brennan, Sox and Burstin (1996). These authors find that iatrogenic injuries in New York account for 3.7% of all hospitalizations and negligent iatrogenic injuries account for 1.0% of hospitalizations. Below is their data for the number of people filing suits:
| Cases | Malpractice Suits | % | |
| No adverse Event | 29,952 | 24 | 0.1% |
| Adverse Event | 1,163 | 13 | 1.1% |
| Negligence | 314 | 9 | 2.9% |
| Totals | 31,429 | 46 |
A second report by Studdert, Mello and Brennan (2004) confirm some of the Brennan, et al.’s findings. Citing a Medical Insurance Feasibility Study, 4.6% of all hospitalizations in California involved iatrogenic injury and 0.8% of all hospitalizations involved negligent iatrogenic injuries. These numbers are similar to the 3.7% and 1.0% which Brennan, et. al. estimate.
There are three issues here which I would like to touch on.
The first is that despite conventional wisdom that physician are nearly infallible, 3-5% of all hospitalizations are due to doctor error. One of the greatest risks facing the American medical system is…well…the American medical system. The easiest and most effective way to decrease the error rate is to integrate Information Technology (IT) into the medical field. The Healthcare IT Guy has some good suggestions.
Malpractice suits are not common. Less than 3% of people who receive negligent physician care actually sue. One must note that it is difficult for a patient to determine if negligence has occurred. ‘Do I feel sick because the treatment is not working or is this the doctor’s fault?’
Although only one in a thousand people who receive no medically induced injury sue, these ‘no injury’ cases make up over half of the malpractice caseload.
Brennan, Sox, Burstin (1996) “Relation Between Negligent Adverse Events and the Outcomes of Medical-Malpractice Litigation” New England Journal of Medicine Vol 335 (26).
Nearly since its inception, the field of Economics has used preference relationships, utility functions, and complete rationality as the theoretical basis for nearly all of its major findings. Nevertheless, while the homo oeconomicus may preform well in certain contexts, in others areas the hypothesis of completely rational humans has been found to be patently false. In a paper by Kenning and Plassman (”NeuroEconomics: An overview from an economic perspective“), the authors show how neurobiology and behavioral economics have formed a recent partnership to better help us understand why people behave the way that they do. Scientists use tools such as electroencephalography (EEG), magnetoencephalography (MEG), positron emission tomography (PET), and functional magnetic resonance tomography (fMRI) to measure brain activity in the context of preference and utility evaluation. NeuroEconomics also examines issues such as trust, fairness, learning and altruism.
An example of how neurology and economics can work together involves the ultimatum game. This game has two players. The first makes an offer to the second of how to split $20 and the second person can accept or reject the offer. If the second person accepts, they split the money according to the first person’s proposal. If the second player rejects, both players get nothing.
Economic theory would tell us that homo oeconomicus should offer a minimal amount to the second player. In experimental studies, however, the most frequent outcome is that the first person makes a ‘fair’ 50/50 offer. NeuroEconomics has found that the anterior insula portion of the brain responds negatively to unfair offers. On the other hand, the dorsolateral prefrontal cortex (DLPFC) portion of the brain involves cognitive processes directed at goal maintenance and the accumulation of as much money as possible. Thus, the person making the offer in the ultimatum game must balance the two desires of their brain.
Economics is a wonderful pursuit which often explains the how of a situation; NeuroEconomics may one day help us answer the question of why.
In November 2004, California passed Proposition 71 which allocated $3 billion over ten years to stem cell research. Many of my friends–especially those in the medical field–strongly supported the bill. While I support stem cell research, I do not approve of the referendum system by which this bill was enacted. Spending $3 billion on unproven research is excessive, and shows how referenda written by small interest groups will lead to extreme outcomes not advocated by the median voter. The average voter may approve stem cell research but I doubt voters would agree to pay $100 per person just for stem cell research, when other programs such as vaccinations, Medicaid, etc. may warrant more money as well.
Today’s L. A. Times (”Faith in ‘Miracle Cures’ Is Fading in South Korea“) casts more doubt on whether the $3 billion of taxpayer money was well spent. South Korea’s claims of stem cells miraculously healing patients have been found be at best overblown and at worst false. Some comments from South Korean stem cell patients:
Do I think that stem cell research is futile? Of course not. The lessons to be taken away from this are:
Medpage Today recently released some statistics regarding the prevalence of smoking cessation among Germans with serious health diseases (”Smokers Can’t Say No in the Face of Serious Circulatory Disease“).
“…the investigators found that among those with a single disorder, the proportion of current smokers among ever-smokers ranged from 29% for myocardial infarction to 44.4% for hypertension.
In addition, unlike smokers with central disease, those with a single peripheral disease ailment were unlikely to quit [point estimate 99%]“
It has been been shown overwhelmingly that smoking is a major cause of many diseases. Why do people continue to smoke in the face of these potential dangers? Dr. Ulrich John, the director of the study, stated that many smokers are uninformed about the relationship between smoking and their disease.
Economists would find this claim dubious. The health dangers of smoking are well known and economists generally assume that individuals are rational. One explanation for this behavior could be hyperbolic discounting of the variety proposed by Laibson (1997) (”Golden Eggs and Hyperbolic Discounting“).
In hyperbolic discounting, individuals rationally choose the amount of smoking they want to have each period using their typical discount rate *. Hyperbolic discounting, however, values the future less than the present. This results in time inconsistent choices. Today I may say that I want to smoke, but will stop in the future. Nevertheless, once I reach the future date specificied, I will value smoking more than I had anticipated. In its original context, Laibson used hyperbolic discounting to explain the phenomenon of Christmas clubs which compel individuals to save. Having smokers post a bond which they would lose if they smoked at a future date is a solution many individuals use to quit. For instance, for every cigarette one has, the individual may compel themselves give a friend a quarter.
* Individuals maximize the following utility function each period: E[u(c_t) + b{SUM}_{j=1 to T} d^j u(c_{t+j})…d is the discount rate; b is the rate of preference between the future and now.
A RAND study published in the American Journal of Managed Care (”Varying Pharmacy Benefits With Clinical Status: The Case of Cholesterol-lowering Therapy“) claims that managed care administrators may be able to vary pharmaceutical co-payment amounts by risk group in order to reduce cost and improve health outcomes. Thus study looks at Cholesterol-Lowering pharmaceuticals.
Co-payments are both a blessing and a curse. If a policy-maker’s goal is to reduce the cost of health care, drug co-payments work well. Consumers become more price sensitive when faced with an out-of-pocket payment. On the other hand, if the co-payment is sufficiently high so that a patient decides to forgo the medication, this can be more costly for the insurance company; with the medication the patient may need hospitalization or emergency department care at a later date.
The RAND study says that co-payments should be targeted by risk group. High risk people should have a low (or zero) co-payment in order to induce them to take their medication and avoid costly hospitalization. Low risk patients–by definition–have a lower probability of being admitted into the hospital and thus will have to pay a higher co-payment.
RAND Conclusions and my Interpretation:
1. Higher co-payments decrease pharmaceutical usage. This believe this is a robust finding.
2. Patients in each risk group characterized as ‘full compliance’ had better health outcomes than those who did not comply with the doctor’s complete prescription. One would guess that “following the doctor’s orders” would lead to better health outcomes. However, it is likely that the ‘full compliance’ effect is overstated in this study. Patients who follow their doctor’s orders exactly may have a healthier life style, be more risk averse, or may be better able to afford the the cost of the drug. If this is the case, part of the health outcome is due to the fact that those who follow their doctors orders completely are generally healthier people. While the estimates may be exagerated, I do not argue with the fact that following a physicians orders exactly does lead to beneficial outcomes for those patients.
3. Lowering co-payments for high risk groups was cost effective. It seems to be the case Cholesterol Lowering (CL) drugs are a more cost effective manner of treating patients than either hospital or emergency department admission. This finding may not be generalizable for other diseases.
Criticism
One limitation of this study is that the relationships observed here among compliance, copayments, and service use may not reflect true causal effects. For example, patients who develop new comorbid conditions may be reluctant to continue their medications, and they also are more likely to be hospitalized…Poor compliance often can be attributed to perceived ineffectiveness, side effects, high costs, and simple forgetfulness.
On the positive side, this study is very intriguing since it mixes medical effectiveness and cost efficiency analysis. More insurance companies should be conducting these types of evaluations in order to create effective policy.
A handful of studies have shown that people who drink one or two glasses of wine per day generally live longer than people who do not drink or who drink beer. Does wine really increase longevity? An article on the Medpage Today website suggests that Eating Well May Explain Wine Drinkers’ Better Health.
Prior studies most likely suffered from a phenomenon with which economists know intimately: Selection Bias.
Wine is generally an “upper class” drink while beer is thought to be a “lower class” drink. Thus, people who drink wine are on average better educated, have more income, and eat healthier food. These three factors have been shown to be highly correlated with longevity. From the prior studies we conclude that well-educated, affluent individuals who eat healthy live longer and also happen to drink more wine; wine is not the cause of their added longevity.
As for me, I’ll stick with a good Belgian beer such as Chimay or Leffe over a glass of Merlot any day.
