Read the Full Video Transcript

Christopher Wallis: Hello, and thank you for joining us for this UroToday Journal Club discussion. Today, we are talking about a paper recently published in JAMA Oncology entitled, Changes in Prostate-Specific Antigen Testing Relative to The Revised USPSTF Recommendation on Prostate Cancer Screening. I’m Chris Wallis, an Assistant Professor in the Division of Urology at The University of Toronto. With me today is Zachary Klaassen, Assistant Professor in the Division of Urology at The Medical College of Georgia. This is the citation for this recent publication led by Dr. Leapman and colleagues published in JAMA Oncology.

Prostate cancer screening, as many will know, is relatively controversial. There are clear benefits in terms of cancer-related morbidity and mortality that have been documented. However, there are substantial harms of over-diagnosis and overtreatment with associated morbidity. As a result of the relative balance of these, in 2012, the US Preventative Services Task Force recommended a grade D recommendation, which says that there is a moderate to high certainty that the harms of PSA screening outweigh the benefits. Dissemination of this guideline recommendation was associated with substantial decreases in PSA testing and the use of diagnostic prostate biopsy.

However, about five years later, a draft guideline was circulated by the same body which changed the recommendation from grade D to grade C. This means that the USPSTF was now advocating for individualized decision-making for men aged 55 to 69 years to undergo PSA testing. The finalized guideline in May 2018, confirmed this recommendation. There are however concerns regarding the migration of patients to more advanced prostate cancer that was associated with the initial recommendation.

So in this work, the authors sought to assess the association between these changes in the guideline recommendations regarding PSA testing and national rates of PSA testing. This is a retrospective cohort study utilizing the Blue Cross Blue Shield Axis data set, which includes 36 health organizations and approximately one-third of all individuals in the US. The authors categorized the time between January 1st, 2013, and December 31st, 2019 into 42 different discreet two-month periods. In each of these periods, they examined male beneficiaries aged 40 to 89 years with coverage for at least 14 months, including the 12 months prior to the two-month period of interest.  They censored men who had a diagnosis code for prostate cancer.

As the guideline recommendations for PSA testing are each specific, the authors utilized these in their analyses. They first assessed overall PSA testing rates among all men aged 40 to 89 and then used the USPSTF age thresholds, namely age 40 to 54, then the recommended age group of 55 to 69, and then the older age group, 70 to 89. Additionally, they used five-year age categories to assess this with somewhat more granularity. Crude counts of PSA testing performed as well as the testing status of each patient was calculated within each two-month-long period. The authors then assessed age-adjusted rates of PSA testing per hundred person-years.

The absolute and relative changes in PSA testing rates were assessed in the two complete calendar years preceding and following the changes in recommendations. The authors then used an interrupted time series analysis, which assumed an immediate effect of the guidelines without any lag. They, therefore, assessed two discrete events, the publication of the draft guidelines in April 2017 and the publication of the final guidelines in 2018, recommended a grade C recommendation for PSA testing.  LOESS smoothing was used and they found that April 2017 was the more appropriate inflection node. Then they segmented regression to assess for changes in both the level and trend of PSA testing.

Now, I'm going to pass it over to Zach to walk us through the results from this interesting analysis.

Zachary Klaassen: Thanks, Chris. So this first figure is an interesting figure. This looks at the PSA testing rates among these insured patients relative up to April 2017 with the publication of the USPSTF draft guideline and you can see here, so breaking down this by age, the blue patients at the top are 55 to 69 years of age. The patients in this brownish color are 70 to 89 years of age. The dark blue is 40 to 89 years, and the orange is 40 to 54. So to the left of this figure is the timeframe from January 2013 to March 2017, and this is when the USPSTF had a grade D recommendation for PSA screening.  And you can see that there is an inflection point at the time of the draft guideline, particularly in men, 55 to 69 years of age, as you can see right here. Interestingly, there was also an inflection point in the men that were 70 to 89 years of age, a slight inflection for men, 40 to 89 years of age, and relative stability in the 40 to 54 years of age. And then when the final guideline was published, we see a further increase in PSA testing, particularly for the men that were older than 54 years of age.

This table looks at the changes in PSA testing rates among these beneficiaries and the calendar year preceding publication of the revised draft statement on PSA testing for prostate cancer relative to the calendar year of 2019. So this is a pre-period here, as you can see from January to December 2016, and a post-period from January to December 2019. And what you can see here is that there was essentially, looking at these P- values and absolute changes, essentially, every age group had increased testing in this post-period. As you can see here breaking down from 40 to 89, 40 to 54, 55 to 69, and 70 to 89. In particular, there were significant absolute changes with regards to patients that were 60 to 64 years of age, an absolute change of 6.5%, as well as patients that were 70 to 74 at 8.2%. So a take-home summary from the somewhat busy tables is that there were significant increases in PSA testing across all age groups in this post-period compared to the pre-period.

This table looks at the results of the interrupted time series models, which Chris talked about. And the take-home message here is that there was an increase in testing as I've highlighted by the intercept here for all of these age groups as well. So again, 40 to 89, 40 to 54, 55 to 69, and 70 to 89.

This table here is a similar table, but it looks at patients that did not have any testing in the preceding 12 month period. So this is potentially patients that were not necessarily plugged into the system who may be new patients, with new discussions about PSA screening. And again, we see a very similar table with increased PSA testing on all age groups, again for age 40 to 89, 40 to 54, 55 to 69, and 70 to 89.

So with regards to the discussion points from this study, the study showed that the rates of PSA testing among privately insured men increased after changes in 2017 to the USPSTF's prostate cancer screening guideline that elevated the PSA screening recommendation for prostate cancer from a grade D to a grade C recommendation. Of note, for ages 55 to 69, an increase in PSA testing was seen in 12.1%, which is the age group supported for testing by the USPSTF. But it also showed increased PSA testing among men aged 40 to 54, and for those greater than 70 years of age, which is not supported by the USPSTF. Increased rates of PSA testing associated with the 2017 draft statement exist in the context of a sharp decrease as having been previously shown in testing that followed the 2012 statement discouraging screening. Changes in PSA testing were implemented rapidly and were temporarily associated with the publication of the draft guideline more so than the final version published one year later.  Increases in PSA testing offer an opportunity to promote evidence-based practices that can further shift the balance and benefits for patients considering prostate cancer screening.

The authors are quite transparent about their analysis and list several important limitations as you can see here. First, they did not examine changes in associated diagnostic procedures, incident prostate cancer cases, or the stage of diagnosis. Secondly, although Blue Cross Blue Shield is the largest source of commercial healthcare claims, it may not reflect all populations since it's skewed towards younger and more socioeconomically advantaged patients with employment-based insurance. Third, based on the anonymized nature of the administrative claims, this study was unable to assess patients' race or family history of prostate cancer. Fourth, based on the nature of administrative claims, they were unable to account for the existence or quality of shared decision-making associated with PSA testing, which are conditions specified by the updated guideline.  And finally, 2019 was the most recent year for which the Blue Cross Blue Shield data was available, and so this analysis is not taking into account the effect of the COVID-19 pandemic, which we have seen in other publications has led to decreased screening for a variety of cancers, including prostate cancer.

So in conclusion, this national cohort study found that rates of PSA testing increased after changes to the USPSTF draft statement in 2017, reversing the trends seen after earlier guidance against PSA testing in 2012 for all patients. A significant increase in testing was observed among patients aged 55 to 69 years of age, which was the age group specified by the USPSTF. However, significant increases in testing were also observed among aged 40 to 54 and aged 70 to 89.

We thank you very much for your attention today, and we hope you enjoyed this UroToday Journal Club discussion, assessing the impact of PSA testing after the USPSTF 2017 guideline recommendations.