Michael Anne Kyle had just begun her Ph.D. in health policy and management at Harvard Business School when her friends started complaining about the health care system.
Many of them were having children for the first time, and they’d never been to the doctor so much in their lives. They’d text Kyle, who is also an ICU nurse, with gripes. “I missed a whole day of work for a ten-minute visit,” they’d say, or, “I spent all this time on hold. Is this normal?” Once she started asking, it seemed everyone had a story.
In 2022, the US spent more than $4 trillion on health care, about 17% of its GDP. In 1960, this fraction was 5%; by 2032, it is predicted to rise to about 20%.
More insidious than these financial burdens on patients are the costs without an obvious price tag — phone calls and faxes, forms and coordination between providers, all of which take time and can be challenging to navigate, and can lead patients to delay or forgo care entirely. And though the administrative complexity of health care is familiar to many, it is not well studied. That’s in part because it can be challenging to document, and in part, Kyle says, because people often accept it as a hassle that will work itself out eventually.
Kyle, who was recently named a STAT Wunderkind, spent her Ph.D. and a postdoctoral research fellowship systematically documenting how the non-financial administrative burdens of health care affect patients. She continues to bring them to light as an assistant professor at the Perelman School of Medicine at the University of Pennsylvania.
“I think it matters a lot to patients. I think it affects them more than we realize,” Kyle said. “And I think there’s an opportunity for us to do better.”
The costs of systemic failures
When Kyle first started working as a nurse in 2006, she quickly noticed systemic failures in the health care system. There were still “never events” — errors that should never happen, such as surgery on the wrong body part, or a patient developing bed sores, or a catheter infection in the hospital. There were payment structures aimed at reducing these issues, but they introduced change through physicians even though nurses were often more responsible for avoiding potential “never events.”
And then there were the patients for whom help came too late.
“So much of the care that you do [in a hospital] is for people who had problems at home that should have been solved days, months, weeks, and years ago,” Kyle said.
Patients with advanced heart failure might visit the doctor for the first time in 20 years, so sick they couldn’t leave the hospital without a new heart. Others had rheumatic heart disease, scarring of the heart caused by childhood illnesses that are treatable with antibiotics or preventable with vaccines.
Kyle wondered if she could reach patients sooner and be part of more systemic change. After five years working in community health organizations to help people enroll in programs such as Medicaid, she began studying health policy.
When her friends started bringing her their health care complaints, Kyle combed through the literature, trying to find research on how administrative burdens kept patients from receiving the best care. When she didn’t find much — in part because standard administrative datasets such as Medicare claims don’t capture these statistics — she began designing a survey that would both document administrative burdens and lay the groundwork for interviews with the patients experiencing them.
“I think everybody understands that patients are frustrated with the health care system,” said Michael Chernew, an economist at Harvard Medical School. “I just don’t think many people want to take the time and effort to study it, where Michael Anne did.”
Together with Austin Frakt, a health economist with appointments at Harvard, Boston University, and the Department of Veterans Affairs, Kyle surveyed 4,000 insured patients as part of the Health Reform Monitoring Survey. About one-quarter of the respondents had delayed or forgone care as a result of administrative tasks such as scheduling appointments and resolving billing problems. (This burden disproportionately fell on people with disabilities.) Kyle and Frakt’s findings showed that administrative burden was affecting patients roughly as often as high financial costs.
Their paper, published in 2021, struck a nerve. A journalist wrote an op-ed for Teen Vogue about how she and others with chronic illnesses spent countless hours navigating such issues. “I had known from social media that I wasn’t alone in my experience, but Kyle and Frakt’s study drove home the insidiousness of the issue,” she wrote. “And, unlike a Twitter thread I wrote in frustration, the findings in their study were quantifiable and academic.”
To Nancy Keating, a professor of health care policy and medicine at Harvard and a primary care physician at Brigham and Women’s Hospital, the survey put a number to a trend she sees in her own practice.
“I think a lot of people just give up, which is also a problem because sometimes you don’t even realize that a patient didn’t start a treatment or medication because they couldn’t get it scheduled,” she said.
Medical professionals experience these complexities, too. Recently, Keating personally spent 30 minutes begging a phone tree of people to explain why her patient’s prescription wasn’t approved.
“My heart sinks whenever I have to make one of those phone calls, because I know it’s not going to be a five-minute discussion,” she said. As a part-time physician, Keating said she has time to make those phone calls; that’s often not feasible for other doctors who spend all day seeing patients.
“There’s lots and lots of different examples, but it shouldn’t be this hard, for both patients and for the doctors,” she said.
Kyle, who has seen Keating’s pager go off in meetings, knows how hard Keating works to keep patients with complex medical situations out of the hospital.
“Having one person who cares is kind of the linchpin in a lot of this stuff,” Kyle said. “It’s great if you have that, but a system can’t be set up on the basis of someone’s good heart.”
The costs of delayed care
After graduating in 2021, Kyle stayed at Harvard to complete a postdoctoral fellowship studying the burdens of cancer patients, who often have particularly frequent contact with the health care system. With Keating, she turned her attention to some of the strategies insurance companies use to manage health care costs, such as prior authorizations. In principle, prior authorizations — through which insurance companies require health care providers to obtain approval before proceeding with a treatment plan — prevent unnecessary, expensive costs. In practice, however, they can lead to delayed care, and Kyle wanted to quantify their impact.
Using Medicare claims data, she compared prescription fills from patients taking the same oral anticancer drug before and after a new prior authorization policy was introduced — and again found quantifiable delays. Requiring prior authorization both increased the odds by seven-fold that a person would discontinue their medication within the next 120 days, and delayed the next fill of their prescription by almost 10 days. That kind of delay could cause stress and, in some cases, potentially even lead to disease progression, although Kyle and Keating’s study did not examine those effects.
Now at UPenn, Kyle aims to determine exactly how administrative burdens impact care — if they cause people to miss appointments or go to the emergency room more, for example — by linking survey data and medical records. Though there is unlikely to be a single fix, Kyle says she thinks small changes will make a difference. Standardizing health care forms could help, as might a more nuanced prior authorization policy that prevents the introduction of new policies on established treatments, or on drugs with a track record of efficacy.
Research devoted to paperwork and its impact might not seem glamorous or essential to some, Kyle says, but to her it feels like the last mile of health care, a key part of making it easier to be a patient.
“In the health care system, we talk about how we want to help people, and then when you go to use the system, it’s not helpful,” she said. The goal, she said, should be to “use people’s time more respectfully and make the system feel more user-friendly.”