Enjoy the weekend with some interesting reads…
Every year, the Centers for Medicare and Medicaid Services (CMS) conducts a recovery audit. In a recent report, Medicare collected over $797 million in Medicare overpayments in 2011. Where do these overpayments come from?
Recovery Audit Contractors returned $488 million in improper payments to the Medicare Trust Fund in 2011. Although this may seem like a large amount of money, Medicare spending in 2010 was $524 billion. Thus, the recovered funds amount to less than two tenths of a percentage of total Medicare spending.
Russell Hutchinson hosts this week’s roundup of risk-related posts from New Zealand, and includes some interesting items from around the globe.
On The Health Care Blog (THCB), Dr. Vineet Arora argues that being a doctor is not as attractive as it once was. She writes:
After all, why go into this much debt and spend so much time in training if your prospects are not much better? More recently, the New York Times article points out job prospects for radiology trainees are thinning, meaning the well known “ROAD” (Radiology, Ophthalmology, Anesthesiology, and Dermatology) to success may soon become a road to nowhere if there are no jobs.
There in lies the question, why become a doctor? If the answer is to make money or to have an easy life, then you probably need to look for a new profession. With healthcare payment reform, doctors can expect lower salaries as bundled payment and cost cutting measures are instituted. Moreover, the demand for healthcare will go up as more patients have insurance, leading to higher patient volumes and the expectation to see more patients with the same amount of time.
On the one hand, these fiscal pressures are pushing U.S. physician salaries closer to those of other nations. On the other hand, the cost of medical education in most other countries is much lower. Instead, one of Dr. Arora’s colleagues argues that being a nurse practitioner may be a lower stress, less financially risky alternative.
Survival analysis is used in many contexts. Some examples include:
How can you estimate the probability of survival from empirical data? One method is to compute a Kaplan-Meier curve (or Kaplan-Meier estimator). The Kaplan-Meier curve calculate survival as the share of individuals who do not fail (e.g., break, die, find a job, of individuals) before a given time period.
One advantage of the Kaplan-Meier curve is that it can take into account censored data. Such as cases where a study ends before failure (or in the medical case often death), or due to other forms of censoring within the study horizon (e.g., removal of a part before failure, patient attrition).
Kaplan-Meier curves are particularly useful for measuring the effectiveness of medical treatments. In randomized controlled trials (RCTs) comparing drug treatments against a placebo, one can create two Kaplan-Meier curves to compare the survival probabilities the two trial arms. Visually, one can often clearly see if the treatment is more effective than the control.
However, is there a statistical method for determining whether the two survival functions are equivalent? In fact, the answer is yes. It is called the log-rank test.
The Healthcare Economist has worked through an example in Excel to demonstrate how to calculate the survival functions as well as the log-rank test (EXAMPLE).
HT: Linda Staub & Alexandros Gekenidis Seminar in Statistics.
HealthPartners argues that the answer is yes. In a 2013 Health Affairs article, they argue the following:
HealthPartners in Minnesota launched an online clinic called virtuwell in late 2010. After more than 40,000 cases, we report an average $88 lower cost per episode compared with care received in traditional settings, strong indicators of clinical effectiveness, and a 98 percent “would recommend” rating from customers. The possibility of extrapolating such savings to larger volumes of cases is compelling.
Although I believe that there will be some savings from online health clinics, I believe that much of this perceived savings is due to patients sorting. If relatively healthier patients use the online health clinic, then it could be the case that average costs will be lower for those who use the online services simply due to patient sorting. The report does risk adjust for patient comorbidities and other factors.
Risk adjustment, however, is always imperfect. Thus, three confounding factors could bias these estiamtes.
Thus, although online health clinics may save money, extrapolating these savings over the entire population is likely to produce an overestimate.
The demand side of the equation–that patients will demand online access to around the clock medical care–may drive the market (especially for young adults and families) in the future.
(more…)
Joe Paduda of Managed Care Matters has posted an excellent edition of Health Wonk Review on health care cost trends, reform implementation, and motivations.
How can you tell if you have a normal distribution? For instance, assume you have data on the results of a drug relative to a placebo. You know the mean and standard deviation of the data, but that does not necessarily imply that the data is distributed in a normal fashion.
How can you do this test? What if you think your drug trial results follow a uniform distribution? Or what if you are interested in comparing these drug trial results against crazy probability distribution you just invented?
There is a simply way how to do this: use the Kolmogorov-Smirnov test (K-S test). The K-S test compares the empirically observed cumulative density function (CDF) in the data against any CDF of your choosing. The K-S test aims to find the maximum difference between your empirical CDF and the assumed CDF at any point in the empirical CDF’s distribution.
Mathematically, the CDF for your emirical distribution is:
The Kolmogorov-Smirnov statistic for a given CDF F(x) is equal to:
An example of how to calculate the K-S statistic is here.
One can use the Kolmogorov distribution (tables) and the K-S statistic Dn to determine the probability that the empirical distribution matches the assumed CDF F(x).
To perform the Kolmogorov-Smirnov equality-of-distributions test in Stata, one can use the ksmirnov command.
Perhaps the most important paper on the effect of health insurance on health, spending, and access to care was released this week. Although randomized controlled trials are rare for measuring the effect of insurance, the design of the Oregon Medicaid program introduced random variation into who received Medicaid coverage. The authors write:
In 2008, Oregon initiated a limited expansion of its Medicaid program for low-income adults through a lottery drawing of approximately 30,000 names from a waiting list of almost 90,000 persons. Selected adults won the opportunity to apply for Medicaid and to enroll if they met eligibility requirements. This lottery presented an opportunity to study the effects of Medicaid with the use of random assignment.
The best objective, concise summary of the results of this study were written by Ashish Jha at The Health Care Blog. He lists the following six primary conclusions.
Whether or not the results of the study indicate that Medicaid is worth the money certainly depends on ones political affiliation.
Michael Cannon of the Cato institute says there is only one way to interpret the results, (to paraphrase) Medicaid is a failure.
There is no way to spin these results as anything but a rebuke to those who are pushing states to expand Medicaid. The Obama administration has been trying to convince states to throw more than a trillion additional taxpayer dollars at Medicaid by participating in the expansion, when the best-designed research available cannot find any evidence that it improves the physical health of enrollees. The OHIE even studied the most vulnerable part of the Medicaid-expansion population – those below 100 percent of the federal poverty level – yet still found no improvements in physical health.
Cannon uses the results to advocate an end to Medicaid expansions.
Others take the silver lining approach. Mike Miesen writes in The Health Care Blog there were some health improvements (although not of the statistically significant variety).
But these figures should be taken in context: because the sample sizes for the above conditions were relatively small, statistically significant outcomes look more disappointing than actual outcomes. For example, the prevalence of hypertension decreased by 7.16% in the Medicaid covered population – a “[change] that would be considered clinically significant.” A larger study population could have shown this finding to be statistically significant, too.
In addition, the study unearthed a number of encouraging findings. Compared to the control population, Medicaid enrollees were 30% less likely to self-report depression; more likely to utilize health care services, including using more prescription drugs, making more visits to the clinic, and having more cholesterol-level screenings done (which increased by almost 50%); and more likely to report higher levels of self-judged health.
Ezra Klein also puts a positive spin on the article.
The Incidental Economist makes a good point that it is important to control expectations.
What is reasonable to expect? How much does private insurance affect these values [health outcomes]? Do we know? No. There is no RCT of private insurance vs. no insurance.