Thesis Thursday: Wenjia Zhu

On the third Thursday of every month, we speak to a recent graduate about their thesis and their studies. This month’s guest is Dr Wenjia Zhu who has a PhD from Boston University. If you would like to suggest a candidate for an upcoming Thesis Thursday, get in touch.

Health plan innovations and health care costs in the commercial health insurance market
Randall P. Ellis, Thomas G. McGuire, Keith M. Ericson
Repository link

What kinds of ‘innovations’ did you want to look at in your research, and why?

My dissertation investigated health plan “innovations” for cost containment, in which certain features are designed into health insurance contracts to influence how health care is delivered and utilized. While specifics may vary considerably across health plans, recent “innovations” feature two main strategies for constraining health spending. One is a demand-side strategy, which aims to reduce health care utilization through high cost-sharing on the consumer side. Plans using this strategy include “high-deductible” or “consumer-driven” health plans. The other is a supply-side strategy, in which insurers selectively contract with low-cost providers whom consumers have access to, thereby directing consumers to those low-cost providers. Plans employing this strategy include “narrow network” health plans.

Despite an ongoing debate about whether the demand-side or supply-side strategy is more effective at reducing costs, there is little work to guide this debate due to challenges in causal inference, estimation, and measurement. As a result, the question of cost containment through insurance benefit designs remains largely unresolved. To shed light on this debate, I investigated these two strategies using a large, multiple-employer, multiple-insurer panel dataset which allowed me to address various methodological challenges through the use of modern econometrics tools and novel estimation methods.

How easy was it to access the data that you needed to answer your research questions?

The main data for my dissertation research come from the Truven Analytic’s MarketScan® Commercial Claims and Encounters Database, which contains administrative claims of a quarter of the U.S. population insured through their employment. I was fortunate to access this database through the data supplier’s existing contract with Boston University, and the entire process of accessing the data involved low effort overall.

Occasionally I needed to refine my research questions or find alternative approaches because certain pieces of information were not available in this database and were hard to access elsewhere. For example, in Chapter 1, we did not further examine heterogeneity of plan coverage within plan types because detailed premiums or benefit features of health plans were not observed (Ellis and Zhu 2016). In Chapter 3, I sought out an alternative approach in lieu of the maximum likelihood (ML) method when estimating provider network breadth because provider identifiers were not coded consistently across health plans in my data, precluding the reliable construction of one key element in the ML method.

Your PhD research tackled several methodological challenges. Which was the most difficult to overcome?

In the course of my research, I found myself in constant need of estimating models that require controlling for multiple fixed effects, each of high dimension (something we called “high-dimensional fixed effects”). One example is health care utilization models that control for provider, patient, and county fixed effects. In these models, however, estimation often became computationally infeasible in the presence of large sample sizes and unbalanced panel datasets. Traditional approaches to absorbing fixed effects no longer worked, and the models with billions of data points could barely be handled in Stata even though it provides some convenient user-written commands (e.g. REGHDFE).

This motivated me and my coauthors to devote an entire chapter in my dissertation to looking into this issue. In Chapter 2, we developed a new algorithm that estimates models with multiple high-dimensional fixed effects while accommodating such features as unbalanced panels, instrumental variables, and cluster-robust variance estimation. The key to our approach is an iterative process of sequentially absorbing fixed effects based on the Frisch-Waugh-Lovell Theorem. By writing up our algorithm into a SAS macro that does not require all data to reside in core memory, we can handle datasets of essentially any size.

Did you identify any health plan designs that reduced health care costs?

Certainly. My dissertation shows that health plans that manage care – imposing cost-sharing, requiring gatekeepers, or restricting consumer choice of providers – spent much less (on procedures) compared to comprehensive insurance plans that do not have any of these “care management” elements, even after controlling for patient selection into plan types.

On the other hand, we did not find evidence that either of the new health plan “innovations” – high cost-sharing or narrow networks – particularly saved health care costs compared to Preferred Provider Organizations (PPOs) (Ellis and Zhu 2016). One possibility is that incentives to control one aspect of spending create compensating effects in other aspects. For example, although high-deductible/consumer-driven health plans shift cost responsibility from employers to enrollees, they did not reduce health care spending due to higher provider prices and higher coding intensity. Similarly, while narrow network plans reduced treatment utilization, they did so mostly for the less severely ill, creating the offsetting incentive of up-coding by providers on the remaining sicker patients.

Based on your findings, what would be your first recommendation to policymakers?

To improve the effectiveness of health care cost containment, my first recommendation to policymakers would be to design mechanisms to more effectively monitor and reduce service prices.

My dissertation shows that while tremendous efforts have been made by health plans to design mechanisms to manage health care utilization (e.g., through imposing a higher cost-sharing on consumers) and to direct patients to certain providers (e.g., through selective contracting), overall cost containment, if any, has been rather modest due to insufficient price reductions. For example, we found that high-deductible/consumer-driven health plans had significantly higher average procedure prices than PPOs (Ellis and Zhu 2016). Even for narrow network plans in which insurers selectively contract with providers, we did not find evidence that these plans were successful in keeping low-cost providers. Difficulties of keeping prices down may reflect unbalanced bargaining power between insurers and providers, as well as special challenges in consumers price-shopping in the presence of complex insurance contract designs (Brot-Goldberg et al. 2017).

Brent Gibbons’s journal round-up for 9th April 2018

Every Monday our authors provide a round-up of some of the most recently published peer reviewed articles from the field. We don’t cover everything, or even what’s most important – just a few papers that have interested the author. Visit our Resources page for links to more journals or follow the HealthEconBot. If you’d like to write one of our weekly journal round-ups, get in touch.

The effect of Medicaid on management of depression: evidence from the Oregon Health Insurance Experiment. The Milbank Quarterly [PubMed] Published 5th March 2018

For the first journal article of this week’s AHE round-up, I selected a follow-up study on the Oregon health insurance experiment. The Oregon Health Insurance Experiment (OHIE) used a lottery system to expand Medicaid to low-income uninsured adults (and their associated households) who were previously ineligible for coverage. Those interested in being part of the study had to sign up. Individuals were then randomly selected through the lottery, after which individuals needed to take further action to complete enrollment in Medicaid, which included showing that enrollment criteria were satisfied (e.g. income below 100% of poverty line). These details are important because many who were selected for the lottery did not complete enrollment in Medicaid, though being selected through the lottery was associated with a 25 percentage point increase in the probability of having insurance (which the authors confirm was overwhelmingly due to Medicaid and not other insurance). More details on the study and data are publicly available. The OHIE is a seminal study in that it allows researchers to study the effects of having insurance in an experimental design – albeit in the U.S. health care system’s context. The other study that comes to mind is of course the famous RAND health insurance experiment that allowed researchers to study the effects of different levels of health insurance coverage. For the OHIE, the authors importantly point out that it is not necessarily obvious what the impact of having insurance is. While we would expect increases in health care utilization, it is possible that increases in primary care utilization could result in offsetting reductions in other settings (e.g. hospital or emergency department use). Also, while we would expect increases in health as a result of increases in health care use, it is possible that by reducing adverse financial consequences (e.g. of unhealthy behavior), health insurance could discourage investments in health. Medicaid has also been criticized by some as not very good insurance – though there are strong arguments to the contrary. First-year outcomes were detailed in another paper. These included increased health care utilization (across all settings), decreased out-of-pocket medical expenditures, decreased medical debt, improvements in self-reported physical and mental health, and decreased probability of screening positive for depression. In the follow-up paper on management of depression, the authors further explore the causal effect and causal pathway of having Medicaid on depression diagnosis, treatment, and symptoms. Outcomes of interest are the effect of having Medicaid on the prevalence of undiagnosed and untreated depression, the use of depression treatments including medication, and on self-reported depressive symptoms. Where possible, outcomes are examined for those with a prior depression diagnosis and those without. In order to examine the effect of Medicaid insurance (vs. being uninsured), the authors needed to control for the selection bias introduced from uncompleted enrollment into Medicaid. Instrumental variable 2SLS was used with lottery selection as the sole instrument. Local average treatment effects were reported with clustered standard errors on the household. The effect of Medicaid on the management of depression was overwhelmingly positive. For those with no prior depression diagnosis, it increased the chance of receiving a diagnosis and decreased the prevalence of undiagnosed depression (those who scored high on study survey depression instrument but with no official diagnosis). As far as treatment, Medicaid reduced the share of the population with untreated depression, virtually eliminating untreated depression among those with pre-lottery depression. There was a large reduction in unmet need for mental health treatment and an increased share who received specific mental health treatments (i.e. prescription drugs and talk therapy). For self-reported symptoms, Medicaid reduced the overall rate screened for depression symptoms in the post-lottery period. All effects were relatively strong in magnitude, giving an overall convincing picture that Medicaid increased access to treatment, which improved depression symptoms. The biggest limitation of this study is its generalizability. Much of the results were focused on the city of Portland, which may not represent more rural parts of the state. More importantly, this was limited to the state of Oregon for low-income adults who not only expressed interest in signing up, but who were able to follow through to complete enrollment. Other limitations were that the study only looked at the first two years of outcomes and that there was limited information on the types of treatments received.

Tobacco regulation and cost-benefit analysis: how should we value foregone consumer surplus? American Journal of Health Economics [PubMed] [RePEcPublished 23rd January 2018

This second article addresses a very interesting theoretical question in cost-benefit analysis, that has emerged in the context of tobacco regulation. The general question is how should foregone consumer surplus, in the form of reduced smoking, be valued? The history of this particular question in the context of recent FDA efforts to regulate smoking is quite fascinating. I highly recommend reading the article just for this background. In brief, the FDA issued proposed regulations to implement graphic warning labels on cigarettes in 2010 and more recently proposed that cigars and e-cigarettes should also be subject to FDA regulation. In both cases, an economic impact analysis was required and debates ensued on if, and how, foregone consumer surplus should be valued. Economists on both sides weighed-in, some arguing that the FDA should not consider foregone consumer surplus because smoking behavior is irrational, others arguing consumers are perfectly rational and informed and the full consumer surplus should be valued, and still others arguing that some consumer surplus should be counted but there is likely bounded rationality and that it is methodologically unclear how to perform a valuation in such a case. The authors helpfully break down the debate into the following questions: 1) if we assume consumers are fully informed and rational, what is the right approach? 2) are consumers fully informed and rational? and 3) if consumers are not fully informed and rational, what is the right approach? The reason the first question is important is that the FDA was conducting the economic impact analysis by examining health gains and foregone consumer surplus separately. However, if consumers are perfectly rational and informed, their preferences already account for health impacts, meaning that only changes in consumer surplus should be counted. On the second question, the authors explore the literature on smoking behavior to understand “whether consumers are rational in the sense of reflecting stable preferences that fully take into account the available information on current and expected future consequences of current choices.” In general, the literature shows that consumers are pretty well aware of the risks, though they may underestimate the difficulty of quitting. On whether consumers are rational is a much harder question. The authors explore different rational addiction models, including quasi-rational addiction models that take into account more recent developments in behavioral economics, but declare that the literature at this point provides no clear answer and that no empirical test exists to distinguish between rational and quasi-rational models. Without answering whether consumers are fully informed and rational, the authors suggest that welfare analysis – even in the face of bounded rationality – can still use a similar valuation approach to consumer surplus as was recommended for when consumers are fully informed and rational. A series of simple supply and demand curves are presented where there is a biased demand curve (demand under bounded rationality) and an unbiased demand curve (demand where fully informed and rational) and different regulations are illustrated. The implication is that rather than trying to estimate health gains as a result of regulations, what is needed is to understand the amount of demand bias as result of bounded rationality. Foregone consumer surplus can then be appropriately measured. Of course, more research is needed to estimate if, and how much, ‘demand bias’ or bounded rationality exists. The framework of the paper is extremely useful and it pushes health economists to consider advances that have been made in environmental economics to account for bounded rationality in cost-benefit analysis.

2SLS versus 2SRI: appropriate methods for rare outcomes and/or rare exposures. Health Economics [PubMed] Published 26th March 2018

This third paper I will touch on only briefly, but I wanted to include it as it addresses an important methodological topic. The paper explores several alternative instrumental variable estimation techniques for situations when the treatment (exposure) variable is binary, compared to the common 2SLS (two-stage least squares) estimation technique which was developed for a linear setting with continuous endogenous treatments and outcome measures. A more flexible approach, referred to as 2SRI (two-stage residual inclusion) allows for non-linear estimation methods in the first stage (and second stage), including logit or probit estimation methods. As the title suggests, these alternative estimation methods may be particularly useful when treatment (exposure) and/or outcomes are rare (e.g below 5%). Monte Carlo simulations are performed on what the authors term ‘the simplest case’ where the outcome, treatment, and instrument are binary variables and a range of results are considered as the treatment and/or outcome become rarer. Model bias and consistency are assessed in the ability to produce average treatment effects (ATEs) and local average treatment effects (LATEs), comparing the 2SLS, several forms of probit-probit 2SRI models, and a bivariate probit model. Results are that the 2SLS produced biased estimates of the ATE, especially as treatment and outcomes become rarer. The 2SRI models had substantially higher bias than the bivariate probit in producing ATEs (though the bivariate probit requires the assumption of bivariate normality). For LATE, 2SLS always produces consistent estimates, even if the linear probability model produces out of range predictions. Estimates for 2SRI models and the bivariate probit model were biased in producing LATEs. An empirical example was also tested with data on the impact of long-term care insurance on long-term care use. Conclusions are that 2SRI models do not dependably produce unbiased estimates of ATEs. Among the 2SRI models though, there were varying levels of bias and the 2SRI model with generalized residuals appeared to produce the least ATE bias. For more rare treatments and outcomes, the 2SRI model with Anscombe residuals generated the least ATE bias. Results were similar to another simulation study by Chapman and Brooks. The study enhances our understanding of how different instrumental variable estimation methods may function under conditions where treatment and outcome variables have nonlinear distributions and where those same treatments and outcomes are rare. In general, the authors give a cautionary note to say that there is not one perfect estimation method in these types of conditions and that researchers should be aware of the potential pitfalls of different estimation methods.



Sam Watson’s journal round-up for 12th February 2018

Every Monday our authors provide a round-up of some of the most recently published peer reviewed articles from the field. We don’t cover everything, or even what’s most important – just a few papers that have interested the author. Visit our Resources page for links to more journals or follow the HealthEconBot. If you’d like to write one of our weekly journal round-ups, get in touch.

Tuskegee and the health of black men. The Quarterly Journal of Economics [RePEc] Published February 2018

In 1932, a study often considered the most infamous and potentially most unethical in U.S. medical history began. Researchers in Alabama enrolled impoverished black men in a research program designed to examine the effects of syphilis under the guise of receiving government-funded health care. The study was known as the Tuskegee syphilis experiment. For 40 years the research subjects were not informed they had syphilis nor were they treated, even after penicillin was shown to be effective. The study was terminated in 1972 after its details were leaked to the press; numerous men died, 40 wives contracted syphilis, and a number of children were born with congenital syphilis. It is no surprise then that there is distrust among African Americans in the medical system. The aim of this article is to examine whether the distrust engendered by the Tuskegee study could have contributed to the significant differences in health outcomes between black males and other groups. To derive a causal estimate the study makes use of a number of differences: black vs non-black, for obvious reasons; male vs female, since the study targeted males, and also since women were more likely to have had contact with and hence higher trust in the medical system; before vs after; and geographic differences, since proximity to the location of the study may be informative about trust in the local health care facilities. A wide variety of further checks reinforce the conclusions that the study led to a reduction in health care utilisation among black men of around 20%. The effect is particularly pronounced in those with low education and income. Beyond elucidating the indirect harms caused by this most heinous of studies, it illustrates the importance of trust in mediating the effectiveness of public institutions. Poor reputations caused by negligence and malpractice can spread far and wide – the mid-Staffordshire hospital scandal may be just such an example.

The economic consequences of hospital admissions. American Economic Review [RePEcPublished February 2018

That this paper’s title recalls that of Keynes’s book The Economic Consequences of the Peace is to my mind no mistake. Keynes argued that a generous and equitable post-war settlement was required to ensure peace and economic well-being in Europe. The slow ‘economic privation’ driven by the punitive measures and imposed austerity of the Treaty of Versailles would lead to crisis. Keynes was evidently highly critical of the conference that led to the Treaty and resigned in protest before its end. But what does this have to do with hospital admissions? Using an ‘event study’ approach – in essence regressing the outcome of interest on covariates including indicators of time relative to an event – the paper examines the impact hospital admissions have on a range of economic outcomes. The authors find that for insured non-elderly adults “hospital admissions increase out-of-pocket medical spending, unpaid medical bills, and bankruptcy, and reduce earnings, income, access to credit, and consumer borrowing.” Similarly, they estimate that hospital admissions among this same group are responsible for around 4% of bankruptcies annually. These losses are often not insured, but they note that in a number of European countries the social welfare system does provide assistance for lost wages in the event of hospital admission. Certainly, this could be construed as economic privation brought about by a lack of generosity of the state. Nevertheless, it also reinforces the fact that negative health shocks can have adverse consequences through a person’s life beyond those directly caused by the need for medical care.

Is health care infected by Baumol’s cost disease? Test of a new model. Health Economics [PubMed] [RePEcPublished 9th February 2018

A few years ago we discussed Baumol’s theory of the ‘cost disease’ and an empirical study trying to identify it. In brief, the theory supposes that spending on health care (and other labour-intensive or creative industries) as a proportion of GDP increases, at least in part, because these sectors experience the least productivity growth. Productivity increases the fastest in sectors like manufacturing and remuneration increases as a result. However, this would lead to wages in the most productive sectors outstripping those in the ‘stagnant’ sectors. For example, salaries for doctors would end up being less than those for low-skilled factory work. Wages, therefore, increase in the stagnant sectors despite a lack of productivity growth. The consequence of all this is that as GDP grows, the proportion spent on stagnant sectors increases, but importantly the absolute amount spent on the productive sectors does not decrease. The share of the pie gets bigger but the pie is growing at least as fast, as it were. To test this, this article starts with a theoretic two-sector model to develop some testable predictions. In particular, the authors posit that the cost disease implies: (i) productivity is related to the share of labour in the health sector, and (ii) productivity is related to the ratio of prices in the health and non-health sectors. Using data from 28 OECD countries between 1995 and 2016 as well as further data on US industry group, they find no evidence to support these predictions, nor others generated by their model. One reason for this could be that wages in the last ten years or more have not risen in line with productivity in manufacturing or other ‘productive’ sectors, or that productivity has indeed increased as fast as the rest of the economy in the health care sector. Indeed, we have discussed productivity growth in the health sector in England and Wales previously. The cost disease may well then not be a cause of rising health care costs – nevertheless, health care need is rising and we should still expect costs to rise concordantly.