dripIn the past two posts (here and here), I have been critiquing the latest study making claims about marital status and cancer survival rates. In this final post in the series, I will show you examples of how the New York Times and an editorial in an academic journal gave readers a misleading impression of what the study really did demonstrate.

Misleading Media Reports: The Example of the New York Times

Read any story about this study in the media and you can probably find lots of statements that give misleading impressions. I’m going to pick on the New York Times. In a 10-paragraph article, I had something skeptical to say about 8 of them.

Married cancer patients live longer than single people who have the disease, suggesting that logistical and emotional support from a loved one may be far more critical to cancer care than previously recognized.

Actually, we have no idea whether logistical or emotional support mattered. None of that was measured.

Numerous studies have suggested that married people have better overall health than single people, but those data likely are skewed by the fact that healthy people are more likely to have opportunities to marry.

That skewing is just one of the many problems with such research. Studies comparing the currently married to everyone else, and then making claims about causality, are making claims that are scientifically indefensible. If you want to claim that getting married makes you healthier, then you need to compare all of the people who ever got married to the people who stayed single. To claim that married people are healthier by doing research in which you set aside all the people who got married, hated it, and got divorced, is to engage in cheater science. If you wanted to make a claim about a new drug, you would not rely on studies which set aside all of the people who took it, hated it, and refused to continue taking it.

‘When you have a spouse who is present when the patient is diagnosed, they are an invested party and they are going to more than likely make sure the patient goes to the doctor, that they get the necessary treatments…’’’

That’s an assumption. It was not demonstrated in the study. It may or may not be true.

The study, published in The Journal of Clinical Oncology, found that single patients were 53 percent less likely to receive appropriate therapy than married patients. The finding suggests that maintaining grueling chemotherapy and radiation schedules and taking medication as prescribed is easier for people who have help from a spouse compared with single people who must manage the logistics of cancer treatment on their own.

The authors did not have any data on whether the single people had any help with logistics. The study did not include any measures of maintaining treatment schedules or taking medication. In fact, we do not even know whether doctors prescribed the same treatments to the single patients and the married patients. And if single patients had been prescribed the same treatments and did not make it through them all as often as married patients did, we don’t know whether other factors may have been in play. For example, maybe the unmarried patients ran out of money sooner. Or maybe the unmarried patients were not treated as well as the married patients were (not just medically, but in other ways, too). Again, my claim is not that the proposed interpretation is wrong, but that we just don’t know.

Unmarried cancer patients also were 17 percent more likely to have late-stage cancer at the time of diagnosis, compared with married patients. That suggests that spouses play a role in encouraging patients to see a doctor, while single people may put off doctor visits, resulting in a more advanced cancer by the time they finally seek a diagnosis.

We don’t know whether spouses played a role in encouraging patients to see a doctor. That was not assessed. Again, everyone is just guessing. All of these guesses could be right or they could be wrong – we just don’t know.

The data do not distinguish between same-sex and opposite-sex couples and don’t account for patients who are engaged or living with a partner. Because some of the people labeled as single in the study probably have a committed partner, it’s likely that the findings actually understate the scope of the problem for people who are truly coping with a cancer diagnosis on their own.

This is the typical one-sided skepticism. What about all of the married people who are living separately – perhaps about 7 percent of them?

Notably, men with cancer showed a greater benefit from marriage than did women. That doesn’t mean husbands are not supportive of wives, but instead suggests that single women do a better job of reaching out for social support than do single men, so the gap between single and married women with cancer is not as great as the gap between single and married men with cancer.

Maybe, but again, we don’t know. Reaching out for social support wasn’t measured. We’re still just guessing.

For doctors and hospitals, the data show that being single is an important risk factor for failing to comply with medical treatments…

Actually, the data don’t show that. This is an interpretation of the results, not a statement of the results. It could be true or it could be false.

The Embarrassing Editorial in the Journal of Clinical Oncology

In an editorial in the journal in which the study was published, the writer described the findings and then said that the data were “incontrovertible.” That’s just embarrassing. The study was not a true experiment. Any study with this design results in data that are open to question.

The editorial then goes on to claim that “strikingly, the benefits of marriage are comparable to or greater than anticancer treatment with chemotherapy.” First, as I described in detail previously, we don’t know if this study tells us anything about the supposed “benefits of marriage.” It may be about differences in wealth or in access to health care or in values or in discriminatory practices that favor married patients over single ones.

Second, no data on chemotherapy were included in the research. The authors compared the magnitude of their own findings to the results of other studies in which the effectiveness of chemotherapy was assessed. They were different studies of different people. They may or may not be comparable.

Third, the supposed “benefits” of marriage were not “comparable to or greater than” chemo. For 4 of the 9 cancers that were included in the analyses, chemo was better; for the other 5, “marriage” was.

The editorial has a pity-the-poor-single-people sensibility. We single people are described as “socially disconnected,” “socially isolated,” and “alienated.” We may also have “limited health literacy.” The greater health outcomes that the writer attributes, without qualification, to marriage, is (in his mind) a testament to “the power of human attachment.” Because how could single people ever have human attachments.

The writer’s intent is positive. He wants to encourage more supportive health care for single people. He just goes about it in a way that makes my inner social scientist cringe.

One Good Thing

Several people who sent me links to media reports about this study were impressed by one thing that they noticed. When it came time to spell out the implications for single people, reporters did not say that they should just get married. In so refraining, they were following the lead of the study’s authors, who ended their article by suggesting that what single people need is more social support, both from medical professionals and beyond.

Hospital patient image available from Shutterstock.