Economics

EMR Data Shows Medicaid Gap Widening Between Expansion and Non-Expansion States

In measuring the effects of health insurance coverage expansion as part of our ACAView initiative with Robert Wood Johnson Foundation (RWJF), an important factor to consider is state policy towards Medicaid expansion.

The intention of the Affordable Care Act (ACA) was to expand coverage through two mechanisms: 1) People with moderate incomes could gain coverage through the exchanges, often encouraged by subsidies; and 2) those with lower incomes could gain coverage through an expansion of Medicaid eligibility to include groups that had not traditionally qualified for Medicaid.

For many years, states had widely varying Medicaid eligibility rules, with some states covering only women and their children in need of public aid and low-income people with disabilities. Other states had expanded eligibility to include people at income levels higher than the federal poverty level.

Given the differing Medicaid expansion decisions among states, we examined our data on visits to primary care physicians (PCPs) separately for states with and without Medicaid expansion.

Screen Shot 2014-07-14 at 5.08.40 PM

Figure 1 shows proportions of visits between January 2012 and May 2014 for four groups of adults (18-64): uninsured individuals in Medicaid-expansion states; uninsured individuals in non-Medicaid expansion states; Medicaid beneficiaries in expansion states; and Medicaid beneficiaries in non-expansion states.

Two observations are worth noting:

    1. ACA coverage expansion appears to be widening a pre-existing gap between states that have elected to pursue Medicaid expansion and those that have not. Providers in the Medicaid-expansion states were already seeing higher proportions of Medicaid beneficiaries in 2013. For example, in December of 2013, 12.3% of 18-64 year- old visits to PCPs in expansion states were from Medicaid beneficiaries, compared with 5.9% in non-expansions states, a 6.4 percentage point differential. By May 2013, that difference had expanded to a 9.3 percentage point differential, as the percent of Medicaid visits increased in Medicaid expansion states but held constant in non-expansion states.
    2. The proportion of uninsured fell in both categories, from 4.5% to 3.3% in expansion states and 7.0% to 5.8% in non-expansion states (figures for January through May for both years, respectively).

Screen Shot 2014-07-14 at 5.08.56 PM

Figure 2 expands the Medicaid payer mix analysis to other specialties.

In Medicaid expansion states, all four specialty types showed a substantial increase in the proportion of visits by Medicaid beneficiaries. In contrast, in non-Medicaid expansion states, the proportion of visits by Medicaid beneficiaries decreased for all four specialty groups.

As a result of these changes, by early 2014 PCPs, surgeons, and other specialists in expansion states saw two to three times more adult Medicaid patients (in proportional terms) than in non-expansion states (for example, 15.6% versus 6.3% for PCPs; 11.6% versus 3.1% for surgeons).

For OB-GYN, the ratio between the proportion of visits by Medicaid beneficiaries in the expansion and non-expansion states is much smaller, 19.4% versus 13.4%. This may reflect more generous Medicaid eligibility in non-expansion states for pregnant women compared to other adults.

As we monitor these metrics, a few questions will be of particular interest:

  • Where will the increase in Medicaid volumes in expansion states level off?
  • To what extent is the increase in Medicaid visits driven by established patients who were previously uninsured?
  • What are the effects of increased Medicaid volumes on medical practices?

We will attempt to address these (and other) complex issues throughout the year.

For a better understanding of our goals, methodology, data sample size, and full findings since the inception of the ACAView series, please read our first report, “First Observations Around the Affordable Care Act.”

Livongo’s Post Ad Banner 728*90

12
Leave a Reply

8 Comment threads
4 Thread replies
0 Followers
 
Most reacted comment
Hottest comment thread
8 Comment authors
try minecraft freeSandeep SharmaJosh GrayRobDon't Go Back to Rockville Recent comment authors
newest oldest most voted
try minecraft free
Guest

I like the helpful information you provide on your articles.
I will bookmark your weblog and test again here frequently.
I’m fairly certain I will learn a lot of new stuff proper here!
Best of luck for the following!

Sandeep Sharma
Guest

In states that are expanding Medicaid coverage under the ACA, the data shows an overall increase in adult (18-64) Medicaid beneficiary patient visits. In expansion states, on average, the percentage of Medicaid-covered patients who are being seen by primary care physicians is rising, with Medicaid patients accounting for 12.3% of care in December of 2013 compared with an increased rate of 15.6% in May 2014. Surgeons and other specialists also show increases. Conversely, states that are not expanding Medicaid coverage have seen Medicaid visits remaining flat. These findings indicate that the implementation of the ACA is widening the gap of… Read more »

Perry
Guest
Perry

Rob,

You nailed all the issues I’ve seen with ACA.
No incentive for primary care to open up to Medicaid, or Medicare for that matter. I don’t even think there’s much incentive for physicians to be in the exchanges.
No incentive for more primary care physicians, period, although the administration claims to be Primary Care friendly.
No cost controls.
Bad planning in my opinion.

Rob
Guest

This is interesting from an academic standpoint, but I am with the others on the skepticism about the end-game of all of this: 1. Are there going to be providers available for these people? Most docs avoid Medicaid if at all possible (due to low reimbursements and difficult patients). 2. What impact will this have on the problems accessing primary care physicians? This increases demand. How is anyone dealing with supply? 3. The basic problem I have with the ACA is that it does nothing for cost control. It’s the lack of cost controls on the part of the payors… Read more »

Don't Go Back to Rockville
Guest

Interesting project. I like the idea of mining your EMR data. If you’re serious about moving science forward, you should consider opening up your data set and making it available for download over the Internet. Even if it is anonymized (it will obvioulsy have to be given HIPAA and other considerations) or a limited data set (the practical way to do things). Give it to us to play with. Your users will love you. It would be interesting to see specifically what services these new patients are requiring. It would also be useful to see what the finanical impact is… Read more »

Perry
Guest
Perry

Good points, Vic. I am a physician, but I do Occupational Medicine so do not take Medicaid. My understanding from colleagues is that Medicaid continues to pay poorly and is a lot of hassle. The other problem is that many Medicaid recipients are seen as poor risks due to non-compliance and frequent no-shows for appointments, which I don’t believe is true across the board, but probably fairly frequent nonetheless. The other point you bring up about Medicaid being a temporary solution, it seems to me that if we are not simultaneously making medical care affordable this will become a more… Read more »

Josh Gray
Guest

Thanks for the comment Perry.

Also worth noting that Medicaid plans pay much higher in some states than others. We have a separate work stream looking at whether Medicaid payment levels seem to correlate with patient access.

Vik Khanna
Guest

This is interesting data, and it certainly seems as though Medicaid expansion delivers the promised results of getting more people into the care system. Some of the things I have an interest in going forward: How easy/difficult is it for Medicaid recipients to find caregivers? Given the traditional provider lament that Medicaid is penurious in the extreme, if you fill the pipeline, but don’t expand the provider base by improving reimbursement (because, presumably, getting paid something is better than getting paid nothing), how does that complicate the process of ensuring that people get care in a timely manner? The results… Read more »

Josh Gray
Guest

Good to hear from you Vik We will be trying to infer access issues, for example by looking at how long it takes to book an appointment for new patients. The issue of capacity that you bring up is critical. Later in the year, we will attempt to look at whether there are any changes in who is doing the work, For example, are nurse practitioners and physicians’ assistants performing a higher proportion of care? If so, for what services? I am not familiar enough with Medicaid research to know if we could meaningfully add to what is known about… Read more »

Al Lewis
Guest
Al Lewis

All well and good so far and it’s great that you are totally on top of this, but you are measuring inputs, so far. Next year put on your agenda to measure HCUP outcomes data for the AHRQ list of avoidable admissions, ACA states to declining states. Can’t be just Medicaid because of the increase in enrollments in the blue states. You’d have to add the HCUP categories of unisured and Medicaid together. It might also be interesting to measure commonly overused procedures to see if overtreatment rises. Keep up the good work and thanks for the post…and next year:… Read more »

Josh Gray
Guest

Thanks for the comment Al. there is still quite a bit of work to be done in terms of basic tracking of how the ACA is filtering through the delivery system: new patient volumes; health status of new vs. established patients; patient obligation; delegation analytics, time to schedule an appointment, and so forth. We will have our hands full with that for a while. Time and resources permitting, I would love to be able to look at outcomes in expansion vs. non-expansion states. We provide software and services primarily for ambulatory care, which will make tracking some outcomes (e.g., readmissions)… Read more »

Al Lewis
Guest
Al Lewis

I could do this for ya. I can make HCUP data sing like John Dean. Put a tickler in your file to call me about 10 months after the year closes. Alternatively we could set up the analysis now and develop a baseline and then see what happens down the road. That’s a better study design because we are selecting the parameters ahead of time rather than trying to fit the data to them