Irene Papanicolas, the director of CheSS, argues that it is useless to ask which health system is “best”. She says many comparisons of health systems falsely reduce complex metrics to single causes and ignore important differences between seemingly similar systems. “What is the best health care system?” she asks, “Is the system that has the best population health outcomes, the best cancer care, the best patient experience, the one that spends the least, or that is free at the point of care?”
Papanicolas, who is a professor of health services, policy and practice, points out that Switzerland, Germany and the Netherlands all have private insurance, but perform better than the US on life expectancy and costs. “If you take 10 different health system scores, there’s no one place that’s going to be consistently the best of them all.”
In the face of this complexity, CHESS and its partners are working to frame questions in ways that yield the kind of empirical findings that can inform policy. Where are health care prices rising the fastest? Do certain drug combinations significantly increase the risk of dangerous falls in the elderly? Are fewer new drugs made available when national health authorities balance health benefit with cost?
Research across borders
Public health researchers are accustomed to narrowly tailoring their research questions, sorting through the messy world of data in order to isolate a dependent variable—comparing the same disease, similar patients, and related care interventions. But the many profound differences between national health systems present unique challenges for researchers who want to compare systems that treat different populations made up of different cultures, using different resources, deployed by professionals with different training.
Enrique Bernal-Delgado has collaborated on international comparisons of health data for more than 25 years. Bernal-Delgado founded the Data Science for Health Services and Policy research group at the Institute for Health Sciences in Aragon (IACS), Spain, where he now serves as a senior scientist. He helped pioneer the “federated research” methodology that is at the core of CHESS’s work, in which research “nodes” in different countries can produce findings in response to the same question while maintaining strict national standards of data privacy.
Health data are highly sensitive, which typically prevents researchers from sharing rich data sets across national borders. In the absence of a common data infrastructure and research methodology, researchers are limited in comparing the collected data. But these kinds of stark comparisons are often open to interpretation: even when similar terminology is used, is a “hospitalization” in Sweden comparable to one in the United States?
“So they go to the hospital, how long does it take to do an operation?” says Papanicolas. “How long do they spend in the hospital? Where do they go when they leave the hospital? Do they have rehabilitation? How long do they have to rehabilitate? Where does rehabilitation take place? Is it in an institution? Is he at home? Do they have a primary care physician? How often do they see their primary care doctor in the following year? Do they go back to the hospital? How many times? What drugs are with? Does their drug regimen change?” This kind of conceptual slippage makes it difficult to answer important questions. “What does their care look like over the course of a year?”
In 2018, before joining the Brown faculty, Papanicolas led the formation of the International Collaboration on Costs, Outcomes and Needs in Care (ICCONIC), which now brings together research teams in 12 different countries to meet this research need and shape policy through partnership.
To make rigorous comparisons using richer data from linked sources—including time-dependent variables—required a new methodology, one that met national legal requirements, as well as what Bernal-Delgado calls “layers of interaction”.
“Semantic interaction isn’t just about speaking the same language – we don’t speak the same language!” says Bernal-Delgado. “But we need the concepts to be the same.” This often means working with clinicians to create a rigorously consistent vocabulary. Organizational and technological interoperability ensure that each partner is able to perform data queries in a consistent and accurate manner. “It’s like making a recipe, but you have a different cook for each component of the meal,” he says Liana Woskiewho also collaborates with Papanicolas.
Creating this federated research infrastructure took a whole year, but Papanicolas hopes things will move faster: “I can accept the first project taking a year, if the second takes a month.”
Natural experiments
Every now and then, health systems experience an identical shock that makes the comparison a little easier. During the COVID-19 pandemic, a new virus tested every national health system, yielding very different responses. Countries followed different policies to treat people and contain the spread of the virus, policies shaped by the characteristics and underlying capacities of their health systems. With support from The Health Foundation, a British charity, Papanicolas and her partners are studying whether lockdowns and other public health interventions, such as school closures and stay-at-home orders, disrupt care. But unlike experts comparing Sweden to America, they are comparing similar patients in different systems.
#health #care #systems #learn #School #Public #Health
Image Source : sph.brown.edu
