Total Pageviews

Tuesday, November 20, 2012

A VIEW FROM THE OFFICE



IF PATIENTS KNEW WHAT THEY'RE SUPPOSED TO KNOW,
WOULD THEY WANT WHAT WE OFFER?
 
CONSUMER RATINGS OF 4 PREVENTIVE INTERVENTIONS: The best way to understand the importance of this new study in Annals of Family Medicine(1) is to take the same survey as the report subjects (354 persons of 977 invited, aged 50-70 from 3 GP practices in New Zealand. Four standard clinical interventions are evaluated:  prevention of hip fractures, prevention of breast and colon cancer, and prevention of heart attacks (from treatment of HTN and hyperlipidemia). Here is your Quiz:

Example A: 5,000 men and women aged between 50 and 70 years diagnosed with osteoporosis are given medication for 10 years to reduce the risk of fractures.
        1. From this group, approximately how many hip fractures do you expect would be prevented by taking this medication?
                (a) 1        (b) 5        (c) 50        (d) 100        (e) 500        (f) 1000

Example B: 5,000 men and women aged between 50 and 70 years from the general population are screened for bowel cancer regularly for 10 years.
        2. From this group, approximately how many deaths due to bowel cancer do you expect would be prevented by screening people for bowel cancer?
                (a) 1        (b) 5        (c) 50        (d) 100        (e) 500        (f) 1000

Example C: 5,000 men and women aged between 50 and 70 years from the general population are given medication for 10 years to decrease their risk of dying of cardiovascular disease.
        3. From this group, approximately how many deaths due to cardiovascular disease do you expect would be prevented by taking this medication?
                (a) 1        (b) 5        (c) 50        (d) 100        (e) 500        (f) 1000

Example D: 5,000 women aged between 50 and 70 years from the general population are screened for breast cancer regularly for 10 years.
        4. For this group, approximately how many deaths due to breast cancer do you expect would be prevented by participation in the screening program?
                (a) 1        (b) 5        (c) 50        (d) 100        (e) 500        (f) 1000

    Alright now, go ahead and commit to an answer (single answer for each one). [Tip: You will get credit for the number closest to the actual value whether on the high side or low side.]



Answer Key: "Hip Fracture Prevention: Alendronate achieves a 53% relative risk reduction for hip fracture among those at high risk of fracture. There is little evidence for the benefit of bisphosphonates beyond 5 years of use, but 1 study that examined use up to 10 years found no significant difference in risk between those treated for 5 and 10 years. The average 10-year hip fracture risk for a 60-year old woman has been calculated as 2.3%. If we assume this risk for the hypothetical group of 5,000 people treated for 10 years, then the number of fractures we would expect to avoid would be 54. We considered 50 to be the correct answer, 1 or 5 to be an underestimate, and 100, 500, or 1,000 to be overestimates."
        "Bowel Cancer Screening: Bowel cancer screening using fecal occult blood testing (FOBT) reduces bowel cancer mortality. The absolute reduction in bowel cancer in bowel cancer mortality with FOBT screening is 1 go 2 deaths avoided per 1,000 people screened over 10 years. For the example in the questionnaire, the correct range of number of deaths avoided would therefore be 5 to 10. We considered answers of 5 to be correct, 1 to be an underestimate, and 50,100, 500 or 1,000 to be overestimates."
        "Cardiovascular Disease Prevention: Treatment of hypertension and hyperlipidemia in primary prevention of cardiovascular disease reduces morbidity and mortality. Antihypertensive medications achieve a 13% relative reduction in mortality, and statins achieve relative reductions in mortality of 12% to 17% (absolute reductions of 0.15 to 0.17 deaths per 100 per year). For the example in the questionnaire, the correct range of number of deaths avoided would be 75 to 85. We considered 50 or 100 to be correct answers, 1 or 5 to be underestimates, and 500 or 1,000 to be overestimates."  
        "Breast cancer screening: Estimates of breast cancer mortality reduction with 10 years of screening range from 1 death avoided for every 337 women screened to 1 death avoided for every 2,500. for the example in the questionnaire, the correct range of number of deaths avoided would therefore be 2 to 15. We considered estimates of 1 or 5 as being correct and 50, 100, 500 or 1,000 to be overestimates."

OK. So how did you do? My guess is that the average practicing primary care physician is likely only to get 1 out of the 4 questions correct.

But now comes the real kicker. The authors of this study then asked the same survey respondents for their own opinions--How many deaths would need to be avoided, in their personal opinion, for each intervention to be worthwhile?    
    For hip fracture prevention, 64% of subjects thought you needed to prevent at least 100 fractures over 10 years for the intervention to be worthwhile. This implies that a majority of subjects would decline the intervention if they were informed of the true absolute value. They are simply not impressed with the value of the intervention.
    For bowel cancer prevention, 73% of the subjects thought that you need to prevent at least 100 deaths from bowel cancer, which is significantly greater than the benefit actually achieved.
   For cardiovascular disease prevention, 46% of the subjects thought that you need to prevent at least 500 cardiovascular deaths for the intervention to be worthwhile. Cardiovascular disease prevention, of the 4 interventions analyzed, had the highest number of subjects who estimated a number of deaths that should be avoided correctly for the intervention to be worthwhile--28% of subjects chose either 50 or 100, which are consistent with the actual data.
    For breast cancer prevention, 69% of subjects thought that at least 50 deaths from breast cancer should be avoided over 10 years of screening for the intervention to be worthwhile whereas the actual benefit in deaths avoided is only 1-5.

COMMENT:  For me informed consent is the key issue and is where we fail our patients most often. The information that patients want to know and should know is what is the general probability that any screening intervention will actually benefit them. Below a certain idiosyncratic threshold they are just not interested. This study shows that a substantial majority of patients would not be interested in these 4 common interventions if appropriately provided with the facts. A key part of the problem is that most physicians are as 'innumerate' as patients are and simple don't know what the risks (incidence) and absolute mortality are with and without screening; this has been amply demonstrated by Gil Welch's book, "Overdiagnosed."(2) You can use your own performance on the 4 question survey above to assess the adequacy of your information. If you're not satisfied with your performance, you can rectify it with Dr. Welch's very entertaining book.



REFERENCES:
 
1.  Hudson B et al. Patients' expectations of screening and preventive treatments. Ann Fam Med 2012; 10: 495-502. doi:10.1370/afm.1407. A free copy of the survey instrument is available online at http://annfammed.org/content/10/6/495/suppl/DC1
 
2. Welch GH. Overdiagnosed: Making people sick in the pursuit of health. Beacon Press. Boston. 2011.