With this post, we are pleased to introduce ACAView, a joint initiative between the Robert Wood Johnson Foundation (RWJF) and athenahealth.
2014 marks the launch of the Affordable Care Act’s (ACA) most important coverage expansion provisions, designed to dramatically reduce the number of uninsured Americans. Between now and the end of 2016, millions of individuals are expected to sign up for subsidized insurance coverage through newly established health care exchanges, or marketplaces.
Other tracking initiatives are closely monitoring the number of individuals that sign up for this coverage as well as those that take advantage of expanding Medicaid coverage in some states.
With ACAView, we will take a different approach. We will focus on the provider perspective; more specifically, how the ACA affects the practice patterns and economics of physicians and other care team members around the country. This is also part of a wider effort, Reform by the Numbers, RWJF’s rich source of timely and unique data about the impact of health reform.
ACAView will monitor the impact of coverage expansion on a monthly basis, mining insights from athenahealth’s cloud-based network of more than 50,000 providers and 50 million patients.
athenahealth is a technology and services provider that delivers physicians the tools and support needed to manage the business and clinical aspects of their medical practices. Our cloud-based, centrally hosted software platform provides us with near real-time visibility into practice patterns of physicians around the country.
Our goal is to inform, exchange ideas, and provide a timely, front-row view of how this landmark legislation affects a robust cross-section of providers across the nation. In subsequent reports, we will examine an evolving set of metrics that address a broad range of topics.
We will also share our analyses on the extent to which our providers represent all providers in the US. For more about our data on practices and patients, as well as our preliminary list of metrics, please read our Methodology report.
No Meaningful Change to Date in New Patient Volumes
Among the many unknown questions surrounding coverage expansion is the number of new patients physicians will accommodate. This is a critical issue because one of the goals of health care reform is to allow individuals to form stable physician relationships, rather than seek care in high-acuity settings or forgo care altogether.
If the ACA is working, we would expect physicians to see a higher percentage of new patients over the course of the year. Over the long term, this number should eventually return to historical levels as these new patients become established.
Figure 1 shows the percentage of provider visits accounted for by new patients, for the first two months of 2013 and 2014. These percentages are based on physician practices active on athenahealth’s network before 2011.
In January through March of 2014, new patients accounted for 17% of visits to primary care practices, down slightly from the 17.9% for the same period in 2013.
Other specialties, with the exception of pediatrics, also showed a slight decline in the proportion of new patient visits. In general, we view these differences as too small, or the timeframe too early, to indicate a meaningful change.
The absence of an increase in new patient visits in the first two months of the year does not surprise us. As of April 1, 7.1 million individuals had enrolled in private insurance plans, through either state or federal marketplaces.
This represents approximately 2% of the population (estimates for Medicaid enrollment were not yet available at the time of this report).
Of that 2%, an unknown proportion of these individuals had some form of previous coverage, so they are not necessarily “newly insured.” It will take some time for newly insured patients to locate new physicians, make appointments with them, and receive care.
We look forward to monitoring this metric to identify meaningful changes when and if they occur.
Will there be Noticeable Differences in the Health Status of New Patients?
Another question is whether the new patients that physicians see are sicker than they were in the past. We will examine this issue in greater detail in our next blog post, but provide an early example here.
Figure 2 shows the percentage of new patients with a diagnosis of diabetes from January through March 2014 compared to January through March 2013.
The chart also shows the same statistics for established patients (patients who have visited in the last two years), who have a higher rate of diabetes.For both groups, we see no change to date, but will continue to monitor this and other chronic disease indicators.
The questions of whether the ACA will cause physicians to devote a higher share of their case mix to seeing new patients, or whether new patients will be more likely to have chronic disease (compared to new patients from previous years, or established patients) are just two issues we will consider over the course of the year.
Other questions we will examine include the following:
- How much will new patients owe for their care?
- How will reimbursement levels change for new patients relative to historical patterns?
- How long will new patients have to wait for care from the time an appointment is made?
Our objective with this joint initiative is to answer a broad set of questions as the year progresses, and we welcome your commentary and suggestions along the way. For continued updates, follow Reform by the Numbers, as well as the ACAView tag on the athenahealth CloudView Blog.
You can connect with our researchers through email at ACAView@athenahealth.com. We also invite you to follow Josh Gray, VP of athenaResearch, on Twitter @JoshGray_hit.
Iyue Sung (@IyueSung) is the director of athenaResearch. Josh Gray (@JoshGray_hit) is vice president of athenaResearch.
Categories: Uncategorized
Very soon this web site will be famous amid all blogging and site-building visitors, due to
it’s good content
This is my first time visit at here and i am really happy to read all at alone place.
Well done collaboration and hopefully more interesting data will be released as this is the only way to convert the piles of data (big…) out there to meaningful insights… Maybe even Athena can release some of that data (aggregated/deidentified) or do an hackathon around it…quite a few data scientists out there would like that..
On the data, regardless of the ACA:
I’d be happy to hear what you think on why the data for DM diagnosis shows this pattern of 6.5% vs. 11% (after 2y)… DM prevalence in the USA is 8.3% (ADA)… so is this a population issue? is it the pattern of diagnosis in the EHR data? This can be useful for future analyses as well…
Thanks for the compliment Jim.
We agree that price transparency is worth understanding better. And you’re idea of providing physicians the ability to post prices, if they so choose, is certainly interesting and technically possible.
My opinion is the industry is “not there yet”, though the idea is starting to spread; for example, Surgical Center of Oklahoma’s transparent pricing and CMS’ recent release of reimbursement data.
In the meantime, we’ll share what we can about utilization and changes in reimbursement patterns. Please do continue to read our posts and thank you again.
@Joe F: “we can easily imagine a future in which most patients, especially those who are elderly or in some other way vulnerable, actually have more intense and continual contact with their doctor or doctor’s office — yet actually have fewer “visits”.”
This is very true. The problem is that positive innovation has been suppressed by too much government involvement. Example: email and increased outpatient care just to mention two.
This is outstanding research! Of course, it is focused on utilization after coverage is established.
Would you ever consider research based on price transparency? I am sure that your company can provide your physicians with the ability to post reimbursable prices. With so many more people in high deductible health plans, it would be valuable to understand how price transparency affects utilization.
Thanks for your reply Joe.
We agree that provider-patient “interaction” aren’t necessarily synonymous with “visits”. Providers are certainly evolving in how they engage their patients. And if they do so with technology tied to an EHR (or other HITs), then that surely would provide “contact” metrics beyond the traditional exam room visit.
Excellent! A good and useful piece of tracking for this moment, and for the ways in which we do medicine now.
But it is not the metric that will allow us to follow how practices will change over time. When we put together the rush of new people with the increased medical needs of the baby boom as more of us enter Medicare age and the paucity of new PCPs, it’s clear that the ways doctors deal with patients will have to, and is likely to, evolve. And as it evolves the ways I believe it is going to evolve, “visits” become an increasingly irrelevant statistic — because more and more patient contact will happen through phone calls, apps, and various gadgets. If we imagine this happening, we can easily imagine a future in which most patients, especially those who are elderly or in some other way vulnerable, actually have more intense and continual contact with their doctor or doctor’s office — yet actually have fewer “visits”.
See also my rant on the misuse of Big Data analytics in pursuit of “terrorists.”
http://www.bgladd.com/Total_Information_Awareness/
“Big Data” runs the risk of being 3,000 miles wide and an inch deep (just look at the Oregon dustup). The sqrt(n) “standard error of the mean” problem — i.e., “statistical” significance (an empirical function of sample size beta “power”) vs actionable / clinical significance.
The risk is dependent on the intended use of the data. I worked in credit risk assessment and modeling in a subprime credit card bank for a number of years. We could be “wrong” 99% of the time (false positives and false negatives) as long as the ones we got “right” (customers we booked who didn’t go bad on us and charge off, but instead proved profitable) floated the entire boat (called “CPA” — cost per acquisition; ours was around $100 each). We made increased record profits every year I was there. See my 2003 white paper on our scorecard development methodology.
http://www.bgladd.com/papers/FNBMwhitepaperRiskScoreModelingStudy2003.PDF
Our astutely calibrated scorecard suite beat the pants off FICO. But, financial analytics and “treatments” are one thing.
Getting it wrong with clinical data is potentially quite another matter.
Thanks for the kind words Andy.
I agree entirely. The generation of large, cohesive health data sets streaming from cloud based EMRs opens up whole new avenues for interesting and useful research, characterized by large numbers of observations, standardized metrics,and compressed time frames. I think the coverage expansion provisions of ACA across 2014 – 2016 is going to be a great test case for this kind of work.
I think this is all very important information to have. Good luck with your research!
This initiative shows the ingenuity and fecundity of modern data science. Rarely do researchers have the precise data they’d like to have to investigate and prove their hypotheses. We have to rely on indirect data, whether it’s the expression of certain traits in the genome or the light from distant parts of the sky. Just today, to show the unexpected value that can lie within data, the Boston Globe published an article examining what local doctors spend on expensive medications, basing their results on the just-released Medicare data. The RWJF/athenahealth initiative is exploratory and will evolve. For example, they admit that their providers might not be
representative of the whole population, and they will adjust for that.
Stay tuned!
That is a central question we also have; which answers and/or leads to other questions. For example, are the newly insured burdened with higher deductibles than before, thereby leading to difficulty paying their total out-of-pocket?
You can see a preliminary list of metrics we plan on rolling out in our methodology synopsis: bit.ly/1hDkljy
Thanks for your comment Bubba. The % of new patients actually declined slightly for PCPs and OBs. One hypothesis is newer practices will naturally have a fewer % of new patients over time.
Also, we use visits since that’s the most reliable measurement of activity.
thanks for your comment.
I think the issue you bring up is key. Will physicians open up their panels to see more patients? Will they work harder? Will they see more patients but slightly reduce visits per patient over the course of the year?
Not at all clear to me how this will play out but we hope to at least measure it closely.
Best,
Josh
JoshGray_hit
This is an interesting approach to studying the effect of the ACA.
Re new patient visits, I would think it would depend on how interested/able providers are. Many of the good docs I know have full panels, or are nearly full, so that a new patient appointment is months out.
Thanks for the note Al. I agree that outcomes will of course be critical to measure in the long term.
Our focus this year is on how ACA (specifically coverage expansion) will affect physician practice, which we consider to be a worthy project and one that our data assets can help with.
Josh
JoshGray_hcit
Netizen:
Thank you for your suggestion. Time on documentation and administrative tasks are interesting, and we do measure that explicitly. Doc time is not currently on our list of metrics for this year, but worth considering for the future.
We’ll take a quick look at some point if we have bandwidth.
Best,
Josh
JoshGray_hit
Thanks Vince, means a lot coming from you. We are just getting going on this project, so any input you have would be quite welcome. I think that a change as fundamental as coverage expansion is bound to affect physician practice, and that’s what we are focused on measuring. The challenge now is picking up early signals of developing changes.
Best
Josh
JoshGray_hit
What will be the effect of deductibles and copays with these new plans?
How much of their time are docs spending on documentation and administrative tasks in their EMR in 2014. How does it compare to 2013?
This looks like an excellent way of approaching this measurement element, but I’m afraid it may be the wrong measurement element. It’s not how many people show up — it’s how much did it help? Otherwise it’s like Claude Rains saying he was rounding up twice the usual number of suspects.
There needs to be an index of things that should change if there is good preventive care: ER visits, prevention-sensitive admissions etc. and see if there is an inflection. Outcomes, not participation, is key.
Nice going. I think you’ve identified an important focus area of ACA implementation that isn’t yet getting a lot of attention. I look forward to reading future posts.
V
The near one percent uptick in PCP and OB/GYN stands out — I wonder what these numbers are going to be in two – three weeks?
Is there a way to track APPOINTMENTS not just VISITS?