1730 WHAT IS THE TRUE NUMBER-NEEDED-TO-SCREEN AND TREAT TO SAVE A LIFE WITH PSA SCREENING?

Article Outline

• INTRODUCTION AND OBJECTIVES
• METHODS
• RESULTS
• CONCLUSIONS
• Copyright

Back to Article Outline

INTRODUCTION AND OBJECTIVES 

The European Randomized Study of Screening for Prostate Cancer (ERSPC) reported a 20% mortality reduction (hazard ratio, HR of 0.80) with PSA screening (NEJM 2009; 360: 1320), which increased to 31% when corrected for noncompliance in the screening arm and contamination in the control arm (Eur Urol 2009; 56:584). Based on the intent-to-treat analysis, they estimated a number-needed-to-screen (NNS) of 1410 and number-needed-to-treat (NNT) of 48 to prevent 1 prostate cancer death at a median follow-up time of 9 years (just when the cumulative hazard functions began to diverge). Although NNS and NNT are useful statistics to assess the benefits and harms of an intervention, there are numerous pitfalls in their calculation and interpretation. The objective of our study was to re-examine the effect of varying follow-up times on the NNS and NNT using simulated data extrapolated from the ERSPC report.

Back to Article Outline

METHODS 

Based on previously reported data from the ERSPC, we modeled the cumulative hazard function using a piecewise exponential model. We assumed a constant hazard of 0.0002 for both the screening and control groups for years 1-7 of the trial; whereas, we assumed different constant rates of 0.00062 and 0.00102 for the screening and control groups for years 8 to 12, respectively. Based on this non-proportional hazards assumption, we computed patient survival and cumulative hazard ratios (CHR) over time as a function of the cumulative hazard function. Annualized cancer detection and drop-out rates were also approximated based upon the observed number of individuals at risk in published ERSPC data.

Back to Article Outline

RESULTS 

Figure 1 compares the simulated cumulative hazard functions to published data from the ERSPC. According to our model, the NNS and NNT at 9 years were 1254 and 43, respectively. This corresponds to a CHR of 0.77 similar to the crude HR of 0.80 from the ERSPC report. Subsequently, the NNS decreased from 837 at year 10 to 503 at year 12, and the NNT decreased from 29 to 18.

• 
  • View Large Image
  • Download to PowerPoint

Back to Article Outline

CONCLUSIONS 

Despite the seemingly simplistic nature of estimating NNT, there is widespread misunderstanding of its pitfalls among the urological community, the media, and the general public. With additional follow-up in the ERSPC, if the mortality difference continues to grow, this will lead to a decrease in the NNT.