Unbiased Analysis of Today's Healthcare Issues

Universal Basic Income

Written By: Jason Shafrin - Jul• 21•17

Universal basic income is the idea that all individuals in a society should be guaranteed a minimum income.  The logic behind this approach is one of equity.  Many members of society feel that all individuals are entitled to some basic level of financial well-being regardless of their skills, ability or willingness to work.

Current government approaches often aim to subsidize certain groups based on their family situation (e.g., single mothers), health (e.g., disabled) or current income (e.g., earned income tax credit).  Administering these programs can be costly.  Universal basic income is simple to administer, provides the maximum fairness, and may have fewer work disincentives than current programs.  For instance, an individual may wish to take a job with higher pay but if doing so means they would lose their Medicaid benefit, they may decide not to do so.  In effect, their marginal income tax rate would be very high, even potentially more than 100% once one incorporates the value of non-cash benefits.

Universal basic income does suffer from at least two problems.  The first is that it may be more succeptible to corruption.  Whereas current government programs may require (onorous) paperwork to verify program elibility, part of universal basic income’s cost savings could come from a lower administrative need, which could result in additional corruption.  For instance, people may stop informing the government that their elderly loved ones have died because they may want to continue getting their loved ones basic income check.  Also, raising a community’s basic income could reduce the number of people interested in participating in the labor market.

One particular study by Calnitsky and Latner (2017) looks at exactly this question based on data from the Mincome initiative, where the province of Manitoba instituted a basic income program for a certain town. They authors found:

Would people work less if their basic needs were guaranteed outside the market? Never before or since the Dauphin experiment has a rich country tested a guaranteed annual income at the level of an entire town. A community-level experiment accounts for the fact that people make decisions in a social context, not in isolation. Using hitherto unanalyzed data we find an 11.3 percentage point reduction in labor market participation, and nearly 30 percent of that fall can be attributed to “community context” effects. Additionally, we show that withdrawals were driven disproportionately by young and single-headed households. Participants who provide qualitative explanations for work withdrawals typically cite care work, disability and illness, uneven employment opportunities, or educational investment.


Atul Gawande on AI

Written By: Jason Shafrin - Jul• 19•17

Tyler Cowen has one of his “Conversations” with Atul Gawande.  The interview is interesting throughout.  Below is an excerpt from their discussion on artificial intelligence.

TYLER COWEN: …How far are we from having an AI that is capable of actually doing diagnosis to people? That is, they might speak into a Skype connection, something like Watson would hear what they say, and they would then diagnose the person well enough that this would be a usable form of healthcare? Is that far, close?

ATUL GAWANDE: Massively far. I think it’s one of the hardest things. You want me to tell why?

COWEN: Tell us why, yes.

GAWANDE: OK, the diagnosis process—people imagine what it is, is that people come to you with a crisply defined problem. “I have symptom one, two, three. I have data to add to it, and now give me the answer.”

The reality is, first of all, people come to you often unable to explain what their problem is. “I have pain.” “Where?” “Hmmm. Well, it’s sort of here.” And they’ll point with a hand. “Well, do you mean there under your rib cage, or you mean in your chest, or . . .”

So you have this probing process that is part of it and how they tell the story. Then there’s also how their story had evolved over time, and they often have to put it in their words. It’s more of a narrative than it is a straight set of data. That’s problem one.

IBM Watson put their AI on this problem, and it would never be the problem I would have put them on. The second part of it is that it changes over time, and you’re adding data along the way. You’re integrating it with a little bit about your view of the understanding of the person and their likelihood to even say that something is a major symptom or not.

There is no question that you can augment the human capability. But the idea that you pull out your phone and it would give you the diagnosis—it is still one of the hardest problems in reducing error in medicine, is the fact that we still have a high rate of error, and the sources of the error have to do with the human being rather than the calculation.

COWEN: But say you only get 15 minutes with your doctor, which is pretty common, and as you know, those conversations don’t always run so well. People are intimidated, they forget the right question to ask. You could have three hours talking to something like Watson. Maybe 80 percent of the dialogue is nonsense, but at the end, you apply machine learning.

And keep in mind, the alternative now is that people use Google, which is in a sense the world’s number one doctor. So AI only needs to be better than Google, which is already a form of AI. In that sense, isn’t it just around the corner that it would be a marginal improvement on what we have today?

GAWANDE: Yeah, one is the replacement question. Can I simply have something that will make the diagnosis? And lots of reasons why that’s difficult. But to augment the human capability, absolutely. There already are programs. One example is called Isabel, where the clinician, having elicited all of this information, can simply put the observations into a list. It will allow them to recognize, “OK, fine. You think that what they have is diagnosis one, but here are eight others in rank order of consideration compared to the one that you think.”

There have been plenty of studies, and it’s been around for more than a decade without the need for AI. This is just crunching some basic data to begin with that can add real value. I think the puzzle of it is that you need that capability to integrate information coming from the person interpreted and be able to get it into these kinds of systems. And in many cases, people may be able to do some of that over time for themselves.

Economic Burden of ACPA+ patients with Rheumatoid Arthritis

Written By: Jason Shafrin - Jul• 17•17

That is the topic of my most recent article in the Journal of Managed Care & Specialty Pharmacy along with co-authors Mahlet Gizaw Tebeka, Kwanza Price, Chad Patel, and Kaleb Michaud.  The abstract of the article–titled “The Economic Burden of ACPA-Positive Status Among Patients with Rheumatoid Arthritis“–is below.

BACKGROUND: Anticitrullinated protein antibodies (ACPAs) are serological biomarkers associated with early, rapidly progressing rheumatoid arthritis (RA), including more severe disease and joint damage. ACPA testing has become a routine tool for RA diagnosis and prognosis. Furthermore, treatment efficacy has been shown to vary by ACPA-positive status. However, it is not clear if the economic burden of patients with RA varies by ACPA status.

OBJECTIVE: To determine if the economic burden of RA varies by patient ACPA status.

METHODS: IMS PharMetrics Plus health insurance claims and electronic medical record (EMR) data from 2010-2015 were used to identify patients with incident RA. Patients were aged ≥ 18 years, had ≥ 1 inpatient or ≥ 2 outpatient claims reporting an RA diagnosis code (ICD-9-CM code 714.0), and had an anticyclic citrullinated peptide (anti-CCP; a surrogate of ACPA) antibody test within 6 months of diagnosis. Incident patients were defined as those who had no claims with an RA diagnosis code in the 6 months before the first observed RA diagnosis. The primary outcome of interest was RA-related medical expenditures, defined as the sum of payer- and patient-paid amounts for all claims with an RA diagnosis code. Secondary outcomes included health care utilization metrics such as treatment with a disease-modifying antirheumatic drug (DMARD) and physician visits. Generalized linear regression models were used for each outcome, controlling for ACPA-positive status (defined as anti-CCP ≥ 20 AU/mL), age, sex, and Charlson Comorbidity Index score as explanatory variables.

RESULTS: Of 647,171 patients diagnosed with RA, 89,296 were incident cases, and 47% (n = 42,285) had an anti-CCP test. After restricting this sample to patients with a linked EMR and reported anti-CCP test result, 859 remained, with 24.7% (n = 212) being ACPA-positive. Compared with ACPA-negative patients, adjusted results showed that ACPA-positive patients were more likely to use either conventional (71.2% vs. 49.6%; P < 0.001) or biologic (20.3% vs. 11.8%; P < 0.001) DMARDs during the first year after diagnosis and had more physician visits (5.58 vs. 3.91 times per year; P < 0.001). Annual RA-associated total expenditures were $7,941 for ACPA-positive and $5,243 for ACPA-negative patients (Δ = $2,698; P = 0.002). RA-associated medical expenditures were $4,380 for ACPA-positive and $3,427 for ACPA-negative patients (Δ = $954; P = 0.168), whereas DMARD expenditures were $3,560 and $1,817, respectively (Δ = $1,743; P = 0.001).

CONCLUSIONS: RA-related economic burden is higher for patients who are ACPA-positive compared with those who are ACPA-negative. Providers may wish to inform patients diagnosed with ACPA-positive RA about the likely future disease and economic burden in hopes that both stakeholders can be more proactive in addressing them.

Why Medicaid patient access to physicians is limited.

Written By: Jason Shafrin - Jul• 16•17

In short, the reason is that Medicaid reimbursement rates for providers is too low.  Saurabh Jha, however, explains the point a bit more artistically in his Health Care Blog piece.

Medicaid pays a cardiologist, with years of training, $25-40 for a consultation to manage a complex patient with multiple comorbidities, on polypharmacy, where the cardiologist must indulge in shared decision making and also ensure the patient adheres to statins.  For comparison, my personal trainer charges me $80. There’s no shared decision making – he tells me to do “burpees” and I must abide or face his wrath.

Note that although Medicaid patients do face challenges to timely access to physicians, particularly in certain regions or for certain specialties, a majority of physicians do accept Medicaid.  The Kaiser Family Foundation found that:

About 70% of office-based physicians accept new Medicaid patients, compared to about 85% who accept new patients with private insurance or Medicare.

This finding does vary across specialty.  For instance, psychiatrists are especially likely to not accept Medicaid patients.


Is VBID gaining a foothold?

Written By: Jason Shafrin - Jul• 14•17

The answer is maybe.  Value-based insurance design ties patient cost sharing to the notion of a treatment’s value.  Higher value treatments have lower cost sharing; lower value treatments have higher cost sharing.  The Incidental Economist writes:

In his own practice, Dr. Fendrick feels as if standard insurance is working against him and his patients. “They are deeply concerned about the amount they have to pay out of their own pockets for the things I beg them to do,” he said. “It makes no sense that they pay the same co-payment for a lifesaving drug to treat diabetes or cancer, as for a drug that makes toenail fungus go away.”

This may be changing. The Affordable Care Act includes a V-BID provision, eliminating cost-sharing for more than 100 preventive services, such as vaccinations and cancer screenings. It’s endorsed by four committees of medical experts.

Many large employers and state governments are going further, reducing cost-sharing for high-value care and medications to treat chronic illnesses, like depression and heart disease. This year, the Centers for Medicare and Medicaid Services began a five-year test of value-based design that permits Medicare Advantage plans in seven states to reduce cost-sharing and enhance benefits for enrollees with designated chronic conditions. Bipartisan legislation has been introduced in the House and Senate to expand the program nationwide.

Some treatments are clear blockbusters.  Others are a complete waste of money.  Most treatments, however, are in the middle where higher cost may lead to more patient benefits.  Thus, a key question to be able to implement VBID is to be able to measure value.  That is one thing I am working on through research at the Innovation and Value Initiative.


Written By: Jason Shafrin - Jul• 13•17

Is your state solvent?

Written By: Jason Shafrin - Jul• 11•17

That is the question that a new study by the Mercatus Center attempts to answer with their “Ranking the States by Fiscal Condition 2017 Edition.” They found that the most financially solvent states (from best to worst) are:

  • 1. Florida
  • 2. North Dakota
  • 3. South Dakota
  • 4. Utah
  • 5. Wyoming

Yet all is not perfect in these top 5 states.  The study asserts:

While these top five states are considered fiscally healthy relative to other states because they have significant amounts of cash on hand and relatively low short-term debt obligations, each state, especially Wyoming, faces substantial long-term challenges related to its pension and healthcare benefits systems.

The five least financially solvent states (from worst to best) are:

  • 50. New Jersey
  • 49. Illinois
  • 48. Massachusetts
  • 47. Kentucky
  • 46. Maryland

The problems related to unfunded liabilities is even more extreme for these bottom tier states, plus these states often do not have sufficient cash to cover all short-term liabilities.


Did P4P work among VA Minorities with Hypertension?

Written By: Jason Shafrin - Jul• 11•17


Do pay-for-performance (P4P) programs work?  P4P programs pay paying providers (or health plans) more if they have better outcomes or follow specific best practices. Whether or not they lead to better outcomes has been much debated.  However, many P4P programs implemented in the real world have been evaluated using a pre-post design which could create endogeneity if the date of the P4P implementation is not random.

Study methodology

In contrast, a recent paper by Petersen et al. (2017) uses a cluster randomized controlled trial design to see whether or not a program to incentivize providers based on the quality of care given to patients with hypertension leads to better outcomes.   This particular study looks at the effect of the hypertension P4P program specifically among black hypertensive patients.  Providers were provided training about the JNC 7 hypertension guidelines.

In the study, the trial:

…randomized 12 VA hospital-based primary care clinics to one of four study groups, differentiated by the type of incentive rewarded: (1) physician-level (individual) incentives; (2) practice-level incentives; (3) physician- and practice-level (combined) incentives; and (4) no incentives (control)…To ensure that facilities of the same type would not be concentrated in one arm, randomization was constrained on hospital teaching status, geographic location, participation in the Antihypertensive and Lipid-Lowering Treatment to Prevent Heart Attack Trial (ALLHAT).

Physicians who had better outcomes (in the treatment arms) received higher payments.  These outcomes included:

  • proportion of sampled black hypertensive patients receiving guideline-recommended antihypertensive medications,
  • proportion with controlled blood pressure,
  • proportion with uncontrolled blood pressure who received an appropriate clinical response to an uncontrolled blood pressure (e.g., lifestyle recommendation for stage 1 hypertension or guideline-recommended medication adjustment)

On average, the total additional payment per physician during the study was $2,744.

P4P programs can also lead to gaming.  Providers who want to maximize their bonus payments may focus on treating patients that are more likely to be adherent to treatment recommendations or who are relatively more healthy.  To test this, the authors look at “whether a patient switched providers, panel turnover among physician participants, and visit frequency.”


Using this approach, the authors found that:

The proportion of black patients who achieved blood pressure control or received an appropriate response to uncontrolled blood pressure in the final performance period was 6.3 percent (95 percent confidence interval [CI] 0.8 11.7 percent; p = .03) greater for physicians in the intervention group than for physicians in the control arm…However, after correcting for multiple comparisons (five study outcomes), the significance threshold did not meet the new adjusted significance level of 0.02.

There was no difference between the arms in patient switching rates and it appears that providers did not decrease visit rates for black patients.

In short there is some suggestive that the P4P improved hypertension outcomes.  Of the five outcome measures considered, all improved more in the intervention arm, but only two of these five results had a p-value of 0.05 or lower.

Healthcare Economist’s Discussion

Even if hypertension outcomes improved, however, it is not clear that P4P is optimal.  For instance, providers may increase their effort to treatment hypertension and improve intermediate hypertension outcomes, but may do this at the cost of focusing on a patient’s other diseases.  In particular, consider the case of a patient who has hypertension as well as a rare but more severe disease.  If the patient is requesting care for their rare disease, providers could treat the patient’s top priority at the risk of decreasing their quality score or they could focus on hypertension care but upset patients by not focusing on the patient’s most severe issues of the day.  This multitasking problem is well known in the field of economics, including work by the 2016 Economics Nobel Prize winner Bengt Holmström.

Value-Based Payments and Incentives to Improve Care

Written By: Jason Shafrin - Jul• 09•17

Please check out my latest publication in Value in Health titled “Value-Based Payments and Incentives to Improve Care: A Case Study of Patients with Type 2 Diabetes in Medicare Advantage“. This work is based on research conducted with my co-authors Jesse Sussell, Kata Bognar, Taylor T. Schwartz, John J. Sheehan, Wade Aubry, and Dennis Scanlon, PhD.


To estimate the impact of increased glycated hemoglobin (A1C) monitoring and treatment intensification for patients with type 2 diabetes (T2D) on quality measures and reimbursement within the Medicare Advantage Star (MA Star) program.


The primary endpoint was the share of patients with T2D with adequate A1C control (A1C ≤ 9%). We conducted a simulation of how increased A1C monitoring and treatment intensification affected this end point using data from the National Health and Nutrition Examination Survey and clinical trials. Using the estimated changes in measured A1C levels, we calculated corresponding changes in the plan-level A1C quality measure, overall star rating, and reimbursement.


At baseline, 24.4% of patients with T2D in the average plan had poor A1C control. The share of plans receiving the highest A1C rating increased from 27% at baseline to 49.5% (increased monitoring), 36.2% (intensification), and 57.1% (joint implementation of both interventions). However, overall star ratings increased for only 3.6%, 1.3%, and 4.8% of plans, respectively, by intervention. Projected per-member per-year rebate increases under the MA Star program were $7.71 (monitoring), $2.66 (intensification), and $10.55 (joint implementation).


The simulation showed that increased monitoring and treatment intensification would improve A1C levels; however, the resulting average increases in reimbursement would be small.



Written By: Jason Shafrin - Jul• 06•17