Brent Gibbons’s journal round-up for 30th January 2017

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.

For this week’s round-up, I selected three papers from December’s issue of Health Services Research. I didn’t intend to to limit my selections to one issue of one journal but as I narrowed down my selections from several journals, these three papers stood out.

Treatment effect estimation using nonlinear two-stage instrumental variable estimators: another cautionary note. Health Services Research [PubMed] Published December 2016

This paper by Chapman and Brooks evaluates the properties of a non-linear instrumental variables (IV) estimator called two-stage residual inclusion or 2SRI. 2SRI has been more recently suggested as a consistent estimator of treatment effects under conditions of selection bias and where the dependent variable of the 2nd-stage equation is either binary or otherwise non-linear in its distribution. Terza, Bradford, and Dismuke (2007) and Terza, Basu, and Rathouz (2008) furthermore claimed that 2SRI estimates can produce unbiased estimates not just of local average treatment effects (LATE) but of average treatment effects (ATE). However, Chapman and Brooks question why 2SRI, which is analogous to two-stage least squares (2SLS) when both the first and second stage equations are linear, should not require similar assumptions as in 2SLS when generalizing beyond LATE to ATE. Backing up a step, when estimating treatment effects using observational data, one worry when trying to establish a causal effect is bias due to treatment choice. Where patient characteristics related to treatment choice are unobservable and one or more instruments is available, linear IV estimation (i.e. 2SLS) produces unbiased and consistent estimates of treatment effects for “marginal patients” or compliers. These are the patients whose treatment effects were influenced by the instrument and their treatment effects are termed LATE. But if there is heterogeneity in treatment effects, a case needs to be made that treatment effect heterogeneity is not related to treatment choice in order to generalize to ATE.  Moving to non-linear IV estimation, Chapman and Brooks are skeptical that this case for generalizing LATE to ATE no longer needs to be made with 2SRI. 2SRI, for those not familiar, uses the residual from stage 1 of a two-stage estimator as a variable in the 2nd-stage equation that uses a non-linear estimator for a binary outcome (e.g. probit) or another non-linear estimator (e.g. poisson). The authors produce a simulation that tests the 2SRI properties over varying conditions of uniqueness of the marginal patient population and the strength of the instrument. The uniqueness of the marginal population is defined as the extent of the difference in treatment effects for the marginal population as compared to the general population. For each scenario tested, the bias between the estimated LATE and the true LATE and ATE is calculated. The findings support the authors’ suspicions that 2SRI is subject to biased results when uniqueness is high. In fact, the 2SRI results were only practically unbiased when uniqueness was low, but were biased for both ATE and LATE when uniqueness was high. Having very strong instruments did help reduce bias. In contrast, 2SLS was always practically unbiased for LATE for different scenarios and the authors use these results to caution researchers on using “new” estimation methods without thoroughly understanding their properties. In this case, old 2SLS still outperformed 2SRI even when dependent variables were non-linear in nature.

Testing the replicability of a successful care management program: results from a randomized trial and likely explanations for why impacts did not replicate. Health Services Research [PubMed] Published December 2016

As is widely known, how to rein in U.S. healthcare costs has been a source of much hand-wringing. One promising strategy has been to promote better management of care in particular for persons with chronic illnesses. This includes coordinating care between multiple providers, encouraging patient adherence to care recommendations, and promoting preventative care. The hope was that by managing care for patients with more complex needs, higher cost services such as emergency visits and hospitalizations could be avoided. CMS, the Centers for Medicare and Medicaid Services, funded a demonstration of a number of care management programs to study what models might be successful in improving quality and reducing costs. One program implemented by Health Quality Partners (HQP) for Medicare Fee-For-Service patients was successful in reducing hospitalizations (by 34 percent) and expenditures (by 22 percent) for a select group of patients who were identified as high-risk. The demonstration occurred from 2002 – 2010 and this paper reports results for a second phase of the demonstration where HQP was given additional funding to continue treating only high-risk patients in the years 2010 – 2014. High-risk patients were identified as having a diagnosis of congestive heart failure (CHF), chronic obstructive pulmonary disease (COPD), coronary artery disease (CAD), or diabetes and had a hospitalization in the year prior to enrollment. In essence, phase II of the demonstration for HQP served as a replication of the original demonstration. The HQP care management program was delivered by nurse coordinators who regularly talked with patients and provided coordinated care between primary care physicians and specialists, as well as other services such as medication guidance. All positive results from phase I vanished in phase II and the authors test several hypotheses for why results did not replicate. They find that treatment group patients had similar hospitalization rates between phase I and II, but that control group patients had substantially lower phase II hospitalization rates. Outcome differences between phase I and phase II were risk-adjusted as phase II had an older population with higher severity of illness. The authors also used propensity score re-weighting to further control for differences in phase I and phase II populations. The affordable care act did promote similar care management services through patient-centered medical homes and accountable care organizations that likely contributed to the usual care of control group patients improving. The authors also note that the effectiveness of care management may be sensitive to the complexity of the target population needs. For example, the phase II population was more homebound and was therefore unable to participate in group classes. The big lesson in this paper though is that demonstration results may not replicate for different populations or even different time periods.

A machine learning framework for plan payment risk adjustment. Health Services Research [PubMed] Published December 2016

Since my company has been subsumed under IBM Watson Health, I have been trying to wrap my head around this big data revolution and the potential of technological advances such as artificial intelligence or machine learning. While machine learning has infiltrated other disciplines, it is really just starting to influence health economics, so watch out! This paper by Sherri Rose is a nice introduction into a range of machine learning techniques that she applies to the formulation of plan payment risk adjustments. In insurance systems where patients can choose from a range of insurance plans, there is the problem of adverse selection where some plans may attract an abundance of high risk patients. To control for this, plans (e.g. in the affordable care act marketplaces) with high percentages of high risk consumers get compensated based on a formula that predicts spending based on population characteristics, including diagnoses. Rose says that these formulas are still based on a 1970s framework of linear regression and may benefit from machine learning algorithms. Given that plan payment risk adjustments are essentially predictions, this does seem like a good application. In addition to testing goodness of fit of machine learning algorithms, Rose is interested in whether such techniques can reduce the number of variable inputs. Without going into any detail, insurers have found ways to “game” the system and fewer variable inputs would restrict this activity. Rose introduces a number of concepts in the paper (at least they were new to me) such as ensemble machine learningdiscrete learning frameworks and super learning frameworks. She uses a large private insurance claims dataset and breaks the dataset into what she calls 10 “folds” which allows her to run 5 prediction models, each with its own cross-validation dataset. Aside from one parametric regression model, she uses several penalized regression models, neural net, single-tree, and random forest models. She describes machine learning as aiming to smooth over data in a similar manner to parametric regression but with fewer assumptions and allowing for more flexibility. To reduce the number of variables in models, she applies techniques that limit variables to, for example, just the 10 most influential. She concludes that applying machine learning to plan payment risk adjustment models can increase efficiencies and her results suggest that it is possible to get similar results even with a limited number of variables. It is curious that the parametric model performed as well as or better than many of the different machine learning algorithms. I’ll take that to mean we can continue using our trusted regression methods for at least a few more years.

Credits

Brent Gibbons’s journal round-up for 12th December 2016

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.

As the U.S. moves into a new era with the recent election results, Republicans will have a chance to modify or repeal the Affordable Care Act. The Affordable Care Act (ACA), also called Obamacare, is a comprehensive health reform that was enacted on the 23rd of March, 2010, that helped millions of uninsured individuals and families gain coverage through new private insurance coverage and through expanded Medicaid coverage for those with very low income. The ACA has been nothing short of controversial and has often been at the forefront of partisan divides. The ACA was an attempt to fill the insurance coverage gaps of the patchwork American health insurance system that was built on employer-sponsored insurance (ESI) and a mix of publicly funded programs for various vulnerable subpopulations. The new administration and republican legislators are promising to repeal the law, at least in part, and have suggested plans that will re-emphasize the private insurance model based on ESI. For this reason, the following articles selected for this week’s round-up highlight different aspects of ESI.

The Mental Health Parity and Addiction Equity Act evaluation study: Impact on specialty-behavioral health utilization and expenditures among “carve-out” enrollees. Journal of Health Economics [PubMed] Published December 2016

Behavioral health services have historically been covered at lower levels and with more restrictions by ESI than physical health services. Advocates for behavioral health system reform have pushed for equal coverage of behavioral health services for decades. In 2008, the Mental Health Parity and Addiction Equity Act (MHPAEA) was passed with a fairly comprehensive set of rules for how behavioral health coverage would need to be comparable to medical/surgical coverage, including for ESI. This first article in our round-up examines the impact of this law on utilization and expenditures of behavioral health services in ESI plans. The authors use an individual-level interrupted time series design using panel data with monthly measures of outcomes. Administrative claims and enrollment data are used from a large private insurance company that provides health insurance for a number of large employers in the years 2008 – 2013. A segmented regression analysis is used in order to measure the impact of the law at two different time points, first in 2010 for what is considered a transition year, and then in the 2011 – 2013 period, both compared to the pre-MHPAEA time period, 2008 – 2009. Indicator variables are used for the different periods as well as spline variables to measure the change in level and slope of the time trends, controlling for other explanatory variables. Results suggest that MHPAEA had little effect on utilization and total expenditures, but that out-of-pocket expenditures were shifted from the patient to the health plan. For patients who had positive expenditures, there was a post-MHPAEA level increase in health plan expenditures of $58.03 and a post-MHPAEA level decrease in out-of-pocket expenditure of $21.58, both per-member-per-month. To address worries of confounding time trends, the authors performed several sensitivity analyses, including a difference-in-difference (DID) analysis that used states that already had strict parity legislation as a comparison population. The authors also examined those with a bipolar or schizophrenia disorder to test the hypothesis that impacts may be stronger for individuals with more severe conditions. Sensitivity analyses tended to result in larger p-values. These results, which were examined at the mean, are consistent with reports that the primary change in behavioral health coverage in ESI was the elimination of treatment limits. In addition to using a sensitivity analysis with individuals with bipolar and schizophrenia, it would have been interesting to see impacts for individuals defined as “high-utilizers”. It would also have been nice to see a longer pre-MHPAEA time period since insurers could have adjusted plans prior to the 2010 effective date.

Health plan type variations in spells of health-care treatment. American Journal of Health Economics [RePEcPublished 12th October 2016

Health care costs in the U.S. were roughly 17.8 percent of the GDP in 2015 and attempts to rein in health insurance costs have largely proved elusive. Different private insurance health plans have tried to rein in costs through different plan types that have a mix of supply-side mechanisms and demand-side mechanisms. Two recent plan types that have emerged are exclusive provider organizations (EPOs) and consumer-driven/high-deductible health plans (CDHPs). EPOs use a more narrowly restricted network of providers that agree to lower payments and presumably also deliver quality care while CDHPs give patients broader networks but shift cost-sharing to patients. EPOs therefore are more focused on supply-side mechanisms of cost reduction, while CDHPs emphasize demand-side incentives to reduce costs. Ellis and Zhu use a large ESI claims-based dataset to examine the impact of these two health plan types and to try to answer whether supply-side or demand-side mechanisms of cost reduction are more effective. The authors present an extremely extensive analysis that is really worth reading. They use a technique for modeling periods of care, called treatment “spells” that is a mix of monthly treatment periods and episode-based models of care. Utilization and expenditures are examined in the context of these treatment “spells” for the different health plan types. A 2SLS regression model is used that controls for endogenous plan choice in the first-stage. The predicted probabilities from plan choice are used as an instrument in the second stage along with a number of controls, including risk-adjustment techniques and individual fixed effects. The one drawback in using the predicted probabilities as the sole instrument is it is not possible to perform an exclusion test. The results, however, suggest that neither of the new plan types performs better than a standardly used health plan. EPOs have the lowest overall spending, but are not significantly different than the standard plan type, and CDHPs have 16 percent higher spending than the standard plan type. The CDHPs in particular have not been studied carefully and these results suggest that previous research on CDHPs found cost-savings due to younger and healthier patients and not because of plan type effects. There are also worries with high deductible plans that patients may elect to forgo necessary healthcare services.

The financial burdens of high-deductible plans. Health Affairs [PubMed] Published December 2016

Having discussed the consumer-directed/high deductible health plans, this third journal article looks at the Medical Expenditure Panel Survey (MEPS) data to examine the burden high deductible health plans place on individuals and families with low incomes. High deductible health plans like the CDHPs are increasingly offered. High deductible plans are sometimes paired with the option to use a flexible spending account (FSA). An FSA gives the patient the option to set aside money from her salary or paycheck that can only be used for healthcare costs, with the benefit that the money set aside will not be subject to various income taxes. The benefit of the high deductible plan is supposed to be lower premiums and the possibility of saving money through the FSA, if that option is available. Yet descriptive analyses using MEPS data from 2011 – 2013 from ESI plans show that high deductible plans impose a particularly high burden on individuals with family incomes below 250 percent of the poverty line. Specifically, the authors found that 29.1 percent of individuals with high deductible plans had financial costs exceeding 20 percent of family income, compared to 20.6 percent of individuals with low deductible plans. For individuals with family income greater than 400 percent of the poverty line, financial burden was not different for high deductible plans compared to other plan types. Yet worryingly, individuals with low incomes were just as likely to have high deductible plans as individuals with high incomes.

Credits