Categorizing and measuring harm from medical treatments

When patients need to decide if a medical intervention is right for them, they need clear information about the potential benefits and harms of the treatment. Unfortunately, we have a lot less information about one side of the equation than the other, because rigorous research on the potential harms of medical care is limited. This makes it harder for both clinicians and patients to make informed decisions about treatments, and often causes them to overestimate the benefits and underestimate the harms of medical care.

In August 2019, 32 experts including researchers, clinicians (primary care doctors and specialists), patients, educators, and journalists, met at the Lown Institute to discuss how we move forward in research on overuse harms. The event was a jumping-off point for several health services researchers led by Dr. Deborah Korenstein, chief of general internal medicine at Memorial Sloan Kettering Cancer Center, to develop recommendations for evaluating the harms of health care services. Their paper was recently published in the Journal of General Internal Medicine, and is an important resource for anyone conducting medical research, as well as journals, regulators, and clinicians.

In this paper, the authors list seven categories of harms to consider when evaluating any medical service, as well as data sources and measurement tools that can be used to gather evidence for each type of harm. Here is a brief overview of the categories:

Physical impairment

Physical impairments resulting from medical treatment–such as pain, disease, death–are the most well-known type of harms. These are often reported in clinical outcomes by clinicians and patients. The authors clarify that researchers should not only report the harms directly resulting from a test or treatment, but also include physical harms from subsequent cascade events (follow-up tests and procedures). These are more difficult to measure, but are incredibly important because clinicians may see certain tests as “harmless” where in reality there are significant consequences from cascade events down the line.

Psychological distress

Less well-known than physical harms are the potential psychological harms that people may experience from medical treatments. These include negative emotions, mood symptoms, or more serious mental health conditions such as depression, anxiety, or PTSD. For example, getting a false positive from a screening test can lead to changes in self-perceptions and anxiety, which can be long-lasting. Other procedures such as hysterectomy can impact ones hormones and mood significantly. These harms are more difficult to find because they may not always be reported, but electronic health records, claims data, and disability claims may have this information.

Social disruption

Social disruption refers to interference with relationships, parenting, and social roles that can happen when people are isolated, suffer physical harm, or have to remove themselves from their social circles because of medical treatment. For example, many women have written about the negative impacts of vaginal mesh complications on their ability to connect intimately with their partner. These harms are not usually measured in clinical trials, but demographic and survey data can be helpful to track changes in life events that could be connected to medical treatments.

Disruption in connection to health care

Being exposed to the unnecessarily harm can erode the trust that patients have in their clinicians, and in the health system as a whole. This impact on the clinician-patient relationship can lead patients to avoid seeking help for health care later when they need it. This type of harm is difficult to measure, but may appear when looking at administrative claims, if someone stopped going to the doctor after a harmful episode.


When someone receives a diagnosis, this can impact their sense of self, and could also impact their future interactions with the healthcare system. Not only can certain diagnoses like cancer increase anxiety, but other diagnoses may have stigma attached, which impacts how other healthcare providers view patients, and how patients view themselves. For example, doctors have warned that giving a diagnosis of polycystic ovary syndrome for mild cases may be doing more harm than good, because it creates anxiety about body weight and future fertility but doesn’t really change the course of medical treatment. Patient surveys or focus groups may be useful in capturing the effects of labeling from medical treatment.

Financial impact

In the US, patients bear a substantial burden of paying for health care. While most people want to do everything they can for their health, the financial cost of care can be extremely disruptive, throwing families into debt or even bankruptcy. While a single imaging test may not be so pricey, cascade events and potential complications can cost patients tens of thousands of dollars. Including data on out-of-pocket costs, bankruptcy, and foreclosure can help illuminate potential financial harms of treatment.

Treatment burden

Having a medical procedure or treatment can seem like a full-time job. It takes a lot of time to go to doctors’ visits, getting the proper medications, dealing with insurance, and taking needed rest. For patients with families and jobs, this treatment burden is exhausting. Patients who cannot do both jobs may be labeled as “non-compliant,” but there is also an opportunity cost to spending so much time dealing with health care issues. Data from patient surveys and interviews, as well as absenteeism measures can help researchers evaluate the harm from treatment burden for a specific treatment.

Korenstein et al. break new ground in this piece by identifying different types of harms that all clinicians and researchers should be aware of. We have to be thinking not only about potential physical harms and complications, but the potential harms from being labeled with a diagnosis, losing trust in the medical system, and financial toxicity. Researchers should consider these harms when designing studies, and funders and journals can require that these harms be reported in new research.

“Perhaps most importantly, clinicians and patients making decisions about health services should be mindful of all domains of harms,” the authors write. By acknowledging these harms in conversations, we can have more fully informed decision-making and better patient-centered care.