Dr. Bala Hota, Tendo SVP and Chief Informatics Officer, is trained as an infectious disease physician. He has also studied epidemiology and holds a Master’s Degree in public health. For the last 20 years, he has worked with advanced technology and data to help create better quality measurements that will ultimately result in improved patient outcomes. The quality systems that are in place in US hospitals today are built off an antiquated model—one that relies on information that is not well-correlated with true quality of care and does not empower patients to understand the best pathways for the right care. Over the coming months, Dr. Hota will be authoring a series of blogs on both the current state and the potential future of quality and value in US healthcare.
In my last post, Measuring quality in the US healthcare industry, I discussed current hospital rankings systems, what measures are being used, and how hospitals are scored. In this post, I’ll examine what’s working within the current systems, what could be improved, and how we can begin building better quality measures for patients.
What’s working.
When looking at some of the current models to measure healthcare quality, one positive thing to note is that most current approaches do not create a huge administrative burden. Why? By relying on billing and administrative data, measures like the CMS Hospital Star Ratings and the US News and World Report Best Hospitals measures don’t require manual chart reviews and are based on readily available data. And because those approaches are completely data-driven, the results are more objective than subjective. Additionally, the current systems are scalable nationwide, enabling broad-based measurement that can promote comparisons between hospitals.
The development of these systems was a major innovation in driving conversations about quality at hospitals. In fact, entire departments within hospitals are typically devoted to quality: assessing the current performance of hospital systems, setting goals for quality improvement, tying external measures to internal measures, and developing culture to enhance performance. The saying, “What’s measured must be managed” is highly applicable here. When measurements are in place, targets can actually be set, and processes for improvement can be put into place.
And what’s not working.
There are a number of downsides with the current system of hospital quality measurement. First, because data for Medicare claims is the most readily available data set that can be used, typically only Medicare patients are being measured in CMS and other quality measures, leaving a large percentage of American healthcare quality performance unaccounted for. In addition, the focus of the entire system is on data, so if the data isn’t perfect, the measurements can’t be, either. Although billing data is a convenient and standardized data source, many aspects of clinical care can be missing in these data. If the goal is to have measures that are useful for patients and meet people where they are in their unique clinical journey, billing data will likely not have the right information to be useful. Multiple studies have shown that consumers find existing approaches—both online ratings and established ranking systems—confusing and do not use quality ratings to select care (1–4).
There are a few reasons that documentation is not a good source of data for quality:
● A challenging infrastructure.
EMRs are hard to use, and physicians are overworked. That combination often means that providers are documenting patient data late at night, and the documentation might not be as complete as necessary to tell the entire patient story, but is geared toward billing. In many cases, just the basics are recorded, but nuances are left out. Documentation and inbox messages are two primary drivers of provider burnout (5), and physicians spend hours after work just to complete tasks (6).
● An incomplete picture. The current systems are built on documentation and billing, but those two components don’t necessarily tell the entire patient story. Sometimes the main narrative of the complete patient story is in other parts of the medical record, or not even collected. For example, consider the information being collected by physicians in their clinical notes about their patients but that is being left out of billing codes because it is not billable. Other detailed data that physicians use to diagnose and treat patients—like laboratory tests and medication information—is not used at all.
● No single standard for how to document. Hospitals and providers all document patient data differently. Because of this variation, hospital rankings may be related more to documentation differences than actual care differences. For example, readmission rate reductions have been found to be more attributable to better severity documentation than actual improvements in rates (7, 8). Anecdotally, we have noted that specialists and surgical specialty documentation may be more focused on acute problems and less likely to include chronic conditions than general medicine physicians, leading to incomplete records. Therefore, the differences in quality measures may be due more to documentation variables than actual quality.
Another important issue to consider is that patients are generally looking for care that involves specific services. When considering an elective surgery or diabetes care, patients aren’t thinking about hospitals, they’re looking at specific doctors or specific services. The current systems don’t do a good job of rating those, instead favoring ratings that rank hospitals as a whole. As a result, the ratings become less meaningful for patients than they are for hospital marketing departments. That puts the focus in the wrong place.
Finally, the current ratings systems miss a lot of more objective data that is present in the EMR because it is difficult to use, and there’s more work involved in getting it. I would argue that these data—which includes laboratory testing, medication prescriptions, and doctors’ notes—is richer data that tells a more accurate story of what’s being done for the patient. A physician can leave out a past history of cancer for a patient when it is not the primary problem, but it is still important to consider when creating a personalized experience for a patient. A review of notes, laboratory tests, and medications would likely reveal this clinical history. Instead of using billing data, we ought to go upstream and use real, raw data from the EMR because it’s all there, waiting to generate insights. Beyond these objective data, there is basically no capture of patient-reported outcomes in our current workflows.
Arguably, these are the most critical pieces of information that patients can provide. For example, how is a patient feeling after a procedure? Are they having pain? When did it subside? I believe capturing these data directly from patients in a standard way will drive a new level of individualized care.
Take the example of knee replacements. For a patient having this procedure, the typical measures of success in the US News and CMS ratings—readmission and mortality—while useful, don’t measure the real outcomes of interest, which are function and pain. If I have osteoarthritis and I’m looking for a provider to perform a knee replacement, I’m going to care the most about what level of function and how much pain I will have at multiple time points: after the procedure, several months after, and in the years post procedure. An assessment of the levels of pain and mobility would be valuable measures that would dramatically personalize insights beyond what we are currently assessing.
Considering better measures.
It’s essential for the healthcare industry to improve how patients and caregivers assess whether quality care is being provided. If better approaches existed, patients would be able to find the most high-quality, effective care for their specific needs. Until more patient-focused measures are developed, the current ratings systems are not likely to go away.
In the meantime, my view is that hospitals should adopt a strategy of gaining a thorough understanding of how the US News, CMS Hospital Stars, and other systems work and develop a clear sense of where strengths and weaknesses are in quality for their care. The overall score of a hospital is a combination of accuracy of submitted data and actual care quality. For hospitals not satisfied with their performance, some of the issues will be present in the data, while others will be caused by care gaps and areas of potential improvement. We need to shine a light on that differentiation and make it transparent so the hospital can take action.
In my next post, we’ll explore how socioeconomic status, social determinants of health, and health equity are impacting health outcomes, and how we can improve.
References.
1. Austin JM, Jha AK, Romano PS, Singer SJ, Vogus TJ, Wachter RM, et al. National hospital ratings systems share few common scores and may generate confusion instead of clarity. Health Aff (Millwood). 2015 Mar 1;34(3):423–30.
2. Emmert M, Schlesinger M. Hospital Quality Reporting in the United States: Does Report Card Design and Incorporation of Patient Narrative Comments Affect Hospital Choice? Health Serv Res. 2017 Jun 1;52(3):933–58.
3. Emmert M, Meszmer N, Schlesinger M. A cross-sectional study assessing the association between online ratings and clinical quality of care measures for US hospitals: results from an observational study. BMC Health Serv Res. 2018 Feb 1;18(1):82.
4. Santa J, D’Alessandro P, Foong S. Scoring healthcare [Internet]. [cited 2022 Nov 20]. Available from: http://assets.fiercemarkets.net/public/webinars/pwc/april2013/console_slides.pdf
5. Yan Q, Jiang Z, Harbin Z, Tolbert PH, Davies MG. Exploring the relationship between electronic health records and provider burnout: A systematic review.
6. Journal of the American Medical Informatics Association. 2021 May 1;28(5):1009–21.
7. Arndt BG, Beasley JW, Watkinson MD, Temte JL, Tuan W-J, Sinsky CA, et al. Tethered to the EHR: Primary Care Physician Workload Assessment Using EHR Event Log Data and Time-Motion Observations. The Annals of Family Medicine. 2017 Sep 1;15(5):419–26.
8. Ibrahim AM, Dimick JB, Sinha SS, Hollingsworth JM, Nuliyalu U, Ryan AM. Association of Coded Severity With Readmission Reduction After the Hospital Readmissions Reduction Program. JAMA Internal Medicine. 2018 Feb 1;178(2):290.
9. Ody C, Msall L, Dafny LS, Grabowski DC, Cutler DM. Decreases In Readmissions Credited To Medicare’s Program To Reduce Hospital Readmissions Have Been Overstated. Health Affairs. 2019;38(1):36–43.