Data Disconnect Surveys Don’t Tell the Story on Cancer Screenings

Presidential elections aren’t the only things that interest pollsters.

You may not know this, but periodic nationwide surveys done by respected government and private organizations have a substantial influence on how we assess the success of our health care system and how we direct our financial resources to address real or perceived problems.

But I have a longstanding bias that some of these surveys don’t reflect the state of affairs in the U.S. when it comes to accurately determining how many people in this country actually do what they tell the pollsters they do when it comes to cancer screening.

A research article in published in March in Oncology Nursing Forum, written by two of my colleagues at the American Cancer Society’s Behavioral Research Center, confirms my suspicion: at least in a part of the African-American community served by federally qualified community health centers, what the patients tell the pollsters doesn’t fit with what the medical records reveal.

The researchers interviewed African-American women who attended these health centers, whose primary mission is to provide medical care to underserved populations. For the most part, the women in this study were economically disadvantaged.

The researchers asked a number of questions about whether or not the women had recent or past cancer screenings. Then, the researchers examined the medical records to see if there was actual documentation that the women had in fact had the recommended screenings and how that checked out with the responses they gave in the interviews.

With regard to mammography to detect early breast cancer, for example, 77{06cf2b9696b159f874511d23dbc893eb1ac83014175ed30550cfff22781411e5} of the women who should have had a screening mammogram said they in fact had one sometime in their lives. The chart showed that only 40{06cf2b9696b159f874511d23dbc893eb1ac83014175ed30550cfff22781411e5} had in fact had a mammogram at some time as noted in the medical record. Also, 29{06cf2b9696b159f874511d23dbc893eb1ac83014175ed30550cfff22781411e5} said they had a mammogram within the past year, and 29{06cf2b9696b159f874511d23dbc893eb1ac83014175ed30550cfff22781411e5} said they had a screening mammogram within the past two to five years. The actual numbers were 9{06cf2b9696b159f874511d23dbc893eb1ac83014175ed30550cfff22781411e5} and 26{06cf2b9696b159f874511d23dbc893eb1ac83014175ed30550cfff22781411e5}, respectively.

Similar findings were reported for clinical breast examination, and two of the three intervals questioned regarding pap smears. The numbers were too small to draw particular conclusions regarding fecal occult blood testing.

Relying on Data

This is no trivial matter.

For example, there is data from the Centers for Disease Control and Prevention regarding the number of African-American women who have had a screening mammogram. Many of us rely on that data for our discussions, presentations and planning.

That data — which is among the best available on the topic — says that 49.9{06cf2b9696b159f874511d23dbc893eb1ac83014175ed30550cfff22781411e5} of African-American women age 40 and older have had a mammogram within the past year, and 64.9{06cf2b9696b159f874511d23dbc893eb1ac83014175ed30550cfff22781411e5} have had one in the past two years.

But other studies that have reported on the frequency of screening mammography in African-American women tell a different story. Past studies have shown that a significant number of African-American women who develop breast cancer all too frequently haven’t had a recent mammogram, the telephone poll results notwithstanding. The same situation exists with cigarette smoking.

When you look at the data available for smoking among African-American youth, the numbers are the lowest for any ethnic group. In fact, the most recent information shows about 11{06cf2b9696b159f874511d23dbc893eb1ac83014175ed30550cfff22781411e5} of African-American youths have smoked within the past 30 days. But when you look at smoking in African-American adults, the numbers are much higher. Since most adult smokers start when they are young, there is something in these numbers that has not made sense to me for a long time.

Unfortunately, I think the difference isn’t that we are doing a much better job of stopping smoking among African-American children than we are among others. I think the surveys aren’t catching the right information.

Not all of my colleagues share my concerns about the accuracy of the available information on screening for various cancers. They tell me essentially that the data is the data.

My concern is that the data is what drives a lot of decision-making by various organizations and governments when it comes to directing money to programs to increase cancer screening. High compliance suggests we are doing a great job, so less money is needed. If the numbers were low, then that would highlight a significant need that would require more funding.

Valid Concerns

I recently had a conversation with a researcher who happens to be interested in this very issue. She told me that my concerns were valid. People will tell the pollster on the phone what they think the pollster wants to hear, rather than admit they haven’t had screening. Or, they may misunderstand what screening is and what it is not. She also commented that the discrepancies are much less when the interviews are done face to face.

We do not do a great job in this country of getting people screened for the early detection of cancer, especially for cancers where we know screening makes a difference such as in breast and cervical cancer.

We need reliable data sources that tell us whether or not we are reaching those populations who are historically underserved, and suffer accordingly when their cancer is detected at a later stage when treatments are more difficult, and successful outcomes less certain.

This particular study does have some limitations as noted by the authors, but it lends credence to my concern that we aren’t getting the information we really need to help us address the very real issues we have in delivering adequate, comprehensive, and effective health care in some of our communities.

Until we recognize and deal with this reality, we may continue to reward ourselves for a job well-done when in fact the job isn’t getting done at all.v

Dr. J. Leonard Lichtenfeld is deputy chief medical officer for the national office of the American Cancer Society. He directs the Society’s Cancer Control Science Department, which produces the Society’s widely recognized guidelines for the prevention and early detection of cancer and guidelines for nutrition and physical activity for cancer survivors. Lichtenfeld is also a board-certified medical oncologist and internist who was a practicing physician for nearly 20 years.