Study 'Shows the Way' to Resolving PSA Controversy

Quality-Adjusted Life-Years Are the Key

Nick Mulcahy

August 15, 2012

August 15, 2012 — A new approach to evaluating the benefits and harms of prostate-specific antigen (PSA) testing "shows the way to a resolution of the long-standing problem about screening for prostate cancer," according to an editorial in the August 16 edition of the New England Journal of Medicine.

"This is welcome news," says editorialist Harold Sox, MD, because the conflicting results from major American and European randomized trials of the screening "did not settle the matter."

A big part of the problem has been that any recommendation about testing, like that recently from the US Preventive Services Task Force, is based on an "apples-and-oranges" comparison, says Dr. Sox, who is from the Dartmouth Institute for Health Policy & Clinical Practice, in Hanover, New Hampshire.

The units of measure are different for benefits (cancer deaths averted) compared with those for harms (overtreatment, erectile dysfunction, urinary problems), he points out.

A new study, published in tandem with the editorial, now addresses the "apples-and-oranges problem" by using the same measure — quality-adjusted life-years — to quantify the harms and benefits of screening, Dr. Sox explains.

"Because the authors express harms and benefits in the same units, they could calculate net benefit, an objective measure of the balance of harms and benefits," he writes.

The new study, which is derived from statistical modeling, found that the benefit of prostate cancer screening is "diminished" by the loss of quality-adjusted life-years caused by overdiagnosis and overtreatment.

Nevertheless, there is an average net benefit from screening in terms of quality-adjusted life-years, according to the study authors, who are led by Eveline Heijnsdijk, PhD, of the Erasmus Medical Center in the Netherlands.

New Estimates

Inputting data from the European Randomized Study of Screening for Prostate Cancer (ERSPC) and other sources into their model, Dr. Heijnsdijk and colleagues found that, for 1000 men of all ages who were followed-up for their entire life span, annual screening of those between the ages of 55 and 69 years would result in:

- 9 fewer deaths from prostate cancer (28% mortality reduction);
- 14 fewer men receiving palliative therapy (35% reduction);
- and a total of 73 life-years gained (an average of 8.4 years per prostate cancer death avoided).

However, the total number of quality-adjusted life-years gained from screening in the group was lower (56 years) than the unadjusted (73 years) because of harms that men endure due to screening, they report. Screening would also result in 45 men being overdiagnosed and overtreated.

The authors calculated that to prevent 1 prostate cancer death, 98 men would have to be screened and 5 cancers would have to be detected.

These numbers are "more favorable" than the 1068 and 48 reported in earlier ERSPC results, the authors acknowledge. But their model "predicts long-term effects after a much longer period," they say, which would allow the control group to have more cancers detected and disease-specific deaths to occur.

These results from the new study are not the final word on evaluating prostate cancer screening because even longer follow-up data are needed from the ERSPC and from quality-of-life analyses. Until longer-term data are available, "universal recommendations regarding screening" should not be made, they conclude.

Dr. Sox agrees, but for different reasons. He finds fault in technical details of the modeling and states that, until more work is done to refine this kind of modeling, "guidelines should avoid recommending for or against PSA screening."

But he is a believer in using quality-adjusted life-years as a basis for prostate cancer screening recommendations.

"The study is a model for developing the evidence base for practice guidelines," Dr. Sox says.

More Precise Measures Still Needed

The new study hinges on the concept of quality of life. To determine quality of life, the authors used "utility estimates" for various health states, such as undergoing a screening test or being treated with prostatectomy or terminal illness. They predicted the number of quality-adjusted life-years associated with screening by using these utility estimates, which were obtained from other studies.

The utility estimates ranged from 0 (death or worst imaginable health) to 1 (full health). The authors explained that a utility estimate of 0.99 was used for the screening phase, "because prostate-cancer screening has little effect on short-term health status and anxiety." On the other hand, there was a utility estimate of 0.95 for all men during the period of 1 - 10 years after diagnosis and after receiving radical prostatectomy or radiation therapy; the lower score implies that some men will have negative outcomes from treatment.

These scores are murky stuff, suggests Dr. Sox. "The authors do not document how they established the range of utilities in the model," he points out.

Nevertheless, "the model calculates the years spent in each health state weighted by the utility of that state," Dr. Sox explains, about how the authors arrive at quality-adjusted life-years.

The authors calculated the effect of the various health states with and without screening over the lifetime of 1000 men who were aged 55 to 69 years at the time of screening (or not being screened).

They assumed that 80% of the men would attend annual screening and found that, overall, an average of 16.7 quality-adjusted life-years would be lost from undergoing screening in the 1000-men group. However, because the model projected that 73 years of life would be gained from screening, the average net effect of screening in the group was 56 life-years gained.

Dr. Sox points out that the net effect of screening can be a loss or a gain, depending on the individual. "It is tempting to speculate that most patients will gain," he writes. However, "these data do not support that conclusion because we do not know the number of patients with low utilities for the health states as compared with the number with higher utilities."

Until better calculations of the distribution of health states are established, Dr. Sox believes it is prudent to hold off on any recommendation on universal screening.

This study was supported by grants from the Netherlands Organization for Health Research and Development, Europe Against Cancer, and the fifth and sixth framework program of the European Union; by grants from agencies or health authorities in the participating countries; and by unconditional grants from Beckman Coulter. A complete list of funding organizations is provided in the Supplementary Appendix. Dr. Sox reports being a board member of Informed Medical Decisions Foundation, which makes decision aids including one for prostate cancer screening. Dr. Heijnsdijk reports having a grant/grant pending with Beckman Coulter, makers of a PSA test. Other authors have relationships with industry, as detailed in the paper.

N Engl J Med. 2012;367:595-605, 669-671. Abstract, Editorial


Comments on Medscape are moderated and should be professional in tone and on topic. You must declare any conflicts of interest related to your comments and responses. Please see our Commenting Guide for further information. We reserve the right to remove posts at our sole discretion.