Diane Harper is internationally recognized for her work in the prevention, diagnosis and treatment of human papillomavirus (HPV)-associated diseases, which include many types of cancer. Her health services research and policy work have focused on guideline development and publication, particularly around cancer screening. Here she talks about recent changes to cervical cancer screening recommendations, the need to balance the potential benefits of early detection with the potential harms of overdiagnosis and overtreatment, and the importance of continually revisiting preventive recommendations to consider new evidence, attitudes, behaviors, and the ever-changing health policy landscape.
How did you first get involved in medicine and women’s health in particular?
I’ve always been interested in women’s healthcare, and that in part stems from my mother, who was very ill for many years, and who never seemed to feel that her concerns were heard or that her questions were answered by health professionals. As a patient, she also didn’t seem to be getting the information she needed to make her medical decisions, and just felt generally dismissed. I had always known I wanted to be in science, but probably would have ended up in a very different field like engineering had it not been for my mother’s experience, which motivated me to switch directions into medicine.
What drove your interest in cervical cancer prevention and HPV?
Cervical cancer is one of the most successfully treated cancers if detected early, and the death rate in the U.S. has more than halved in the last three decades. Yet we also recognize that too much screening and treatment can actually cause great harm to women and their current and future reproductive health. Early on in my clinical work, I was seeing large amounts of overdiagnosis, overtreatment, and overspending around cervical cancer, but these problems weren’t being acknowledged within the medical system or health economic policies. I realized that addressing this imbalance could be an important area where I could make a difference.
Earlier in my career, I was honored to help write the guidelines known as the Bethesda System to help standardize how pathologists should interpret Pap smears, which is a test that looks for signs of cancer in cervical cells. I was part of the Steering Committee for ALTS, a National Cancer Institute-funded intramural program from 1996-2000 that determined that we, in the U.S., had been overtreating people with low-grade disease and needed to put an end to that. Our policy group also said that that cervical cancer screening and contraception should not be linked, and that providers should stop requiring Pap smears as a condition of accessing birth control. Our group also stated that we had to stop routinely screening women 18 or younger, which had been common practice, and in 2012 we set the policy age at 21, which is where it is now.
When I began focusing my research on HPV, the NCI was just coming out with evidence about the correlation between HPV and its role in cervical cancers. And we now know this virus causes nearly all cervical cancers.
How have cervical cancer screening recommendations changed recently?
There are two kinds of tests to screen for cervical cancer: a test for the presence of human papillomavirus (HPV), and a Pap smear. These tests are used by themselves or with each other when one is abnormal.
Over the years cervical cancer screening recommendations have been periodically updated to reflect our understanding of the connection between the HPV virus and different types of cancer, and new evidence about potential benefits and harms of screening and treatment.
I was a member of the U.S. Preventative Services Task Force during its recent review of the evidence behind current cervical cancer screening recommendations. In August 2018, the USPSTF for the first time recommended giving women aged 30-65 years the option of getting tested for high-risk types of HPV every 5 years in place of a Pap smear, or combining these screening methods (known as co-testing) every 5 years, or opting for the Pap test every 3 years.
The major change from the previous recommendations is that the USPSTF now recommends standalone HPV testing every 5 years as an alternative to screening every 3 years with Pap smear alone among women in this age group with previously normal tests. The sensitivity of the HPV test is much higher than the sensitivity of a Pap test and can detect populations at risk more accurately. The assurance of a negative high-risk HPV test lasts for at least 5 years, while the reassurance of a normal Pap test only lasts three years. This saves women time in their health maintenance, allows physicians to address other health needs that are higher priority, and is of great value to the health economics of cancer prevention.
The recommendation to begin screening at age 21 comes from evidence showing that routine cervical cancer screening with any method for women younger than that would likely lead to more harm than benefit, since it might result in unnecessary invasive testing, or harms from overtreatment of conditions that would have turned out to be benign.
HPV infections are very common in young people and often resolve on their own, so it doesn’t make sense to detect the nearly 40 percent of the population who have HPV. It does make sense to see if that HPV has caused changes in the cervix that could lead to cancer, which is what the Pap test can offer women 21-29 years old every 3 years as is recommended by the USPSTF. Women 30 years and older have a lower likelihood of having a high-risk HPV infection, which makes the HPV test more accurate than the Pap test for these women.
What’s new with HPV vaccination?
For the last 20 years, I’ve also been involved with work in HPV vaccine development and vaccine administration guidelines. HPV vaccination prevents infections that can lead to cancers and other diseases, which is why vaccination along with screening is an important part of cervical cancer prevention. In October 2018, the FDA approved the expanded use of the HPV vaccine Gardasil9 (Merck) for girls and boys, women and men 9-45 years of age in the U.S. (the previous approval had been for males and females ages 9-26 years).
Although many adults have already been exposed to the HPV virus, it’s generally only to been to one type, while the vaccine prevents against seven different cancer-associated types and two genital wart types. Broadening the vaccine’s age range now gives people in midlife, who may have new or different partners from when they were younger, the opportunity to get vaccinated against HPV for the first time. And the flip side of that, and why HPV vaccination is also so important, is that a woman who tests negative for HPV can be assured nearly 100 percent protection against half of the high-risk HPV types that are associated with cancer if she receives the vaccine.
Now it will be up to the Centers for Disease Control and Prevention (CDC)’s Advisory Committee on Immunization Practices, or ACIP, to make a recommendation on whether everyone in this expanded age range is recommended to get the vaccine, and insurance coverage decisions generally follow these decisions.
What’s involved in developing guidelines around preventive or other healthcare services? How does evidence really get implemented into health policy?
We can’t make guidelines about anything if the data and research evidence are not there to support it, so we have to have good level of evidence behind the recommendations we make, which requires going deep into what the data really show. It’s incredibly important to realize what our gaps in knowledge are, to understand the limitations of available evidence, and to determine priorities for additional research.
With time, the science changes, disease prevalence changes, treatments change, people’s perceptions change, interpretations of policy change, and this requires us to continually reevaluate the evidence behind decisions and recommendations. The only constant in health policy is that nothing is ever really finished, and everything requires frequent revisiting! We also know that it can take 10-15 years, sometimes more, for healthcare to adopt new evidence-based findings – so the ability for practice to keep up with the science is another challenge.
In my past, I was a state-level leader in the CDC’s National Breast and Cervical Cancer Early Detection Program. We recognized that while women who had private insurance or coverage through Medicare or Medicaid were getting screened, there was a pretty big swath remaining – almost a third of the population – who could not afford these services.
Over its lifetime, the program has filled a crucial gap in health policy by providing breast and cervical cancer screening and diagnostic services to more than 4.8 million low-income, underinsured, and uninsured women across the country, reaching women who may not have access to these services otherwise.
Can you describe what’s meant by shared decision-making in healthcare, which is another focus of your work?
A simple definition is that patients and providers are working together to come up with shared expectations for care that balance patient preferences and values with the evidence about expected outcomes. For me, the key to shared decision making is better understanding and measuring how patients view risks and benefits, what is most important to them, and how they prioritize managing a health issue within the broader context of their lives.
Things like getting time off work for appointments, or paying for parking or childcare, or anxiety over visiting a clinic – these kinds of things can be major barriers to effective care, and particularly so if we on the provider side don’t recognize or appreciate them. While it’s incredibly important for people to be involved in their healthcare, actually enabling that through policy and practice can be really hard.
Is there anything about preventive services that you think is not well understood?
There’s an idea that all prevention efforts are universally good ideas, and while there’s some general truth behind that, the picture is much more complex. For one thing, we can’t separate costs from questions about what good decisions around preventive services are, and so we need to do economic modelling that looks at cost analyses, how many dollars must be spent to gain improved health or quality of life in the areas that matter to people – all those things need to be considered together to make sure we’re allocating resources wisely for the best expected outcomes.
There’s also the issue of appropriate use of screening as well as understanding the threshold between detecting and overdiagnosing. Understanding overdiagnosing is still a very young science, and we need to do a better job of role modeling how we determine these boundaries as we train our newest health professionals.