Background

Hemoglobin A1C (HbA1C) is associated with increased risk of cardiovascular events, but its use for prediction of cardiovascular disease (CVD) events in combination with conventional risk factors has not been well defined.

Methods and results

To understand the effect of HbA1C on CVD risk in the context of other CVD risk factors, we analyzed HbA1C and other CVD risk factor measurements in 2000 individuals aged 40 to 79 years without pre-existing diabetes mellitus or CVD from the 2011 to 2012 National Health and Nutrition Examination Surveys survey. The resulting regression model was used to predict the HbA1C distribution based on individual patient characteristics. We then calculated post-test 10-year atherosclerotic CVD risk incorporating the actual versus predicted HbA1C, according to established methods, for a set of example scenarios. Age, sex, race/ethnicity, and traditional cardiovascular risk factors were significant predictors of HbA1C in our model, with the expected HbA1C distribution being significantly higher in non-Hispanic black, non-Hispanic Asian, and Hispanic individuals than that in non-Hispanic white/other individuals. Incorporating the expected HbA1C distribution into pretest atherosclerotic CVD risk has a modest effect on post-test atherosclerotic CVD risk. In the patient examples, we assessed that having an HbA1C of <5.7% reduced post-test risk by 0.4% to 2.0% points, whereas having an HbA1C of ≥6.5% increased post-test risk by 1.0% to 2.5% points, depending on the scenario. The post-test risk increase from having an HbA1C of ≥6.5% tends to approximate the risk increase from being 5 years older.

Conclusions

HbA1C has modest effects on predicted atherosclerotic CVD risk when considered in the context of conventional risk factors.

Background

Coronary heart disease (CHD) is the leading cause of death in the United States. Previous research examining physicians ability to estimate cardiovascular risk has shown that physicians' generally overestimate the absolute risk of CHD events. This question has, however, only studied risk prediction for a limited number of patient care scenarios. The aim of this study is to measure the ability of physicians to estimate the risk of CHD events in patients with no previous history of coronary heart disease.

Methods

Twelve primary prevention scenarios with a 5-year risk of CHD events were developed. This questionnaire was surveyed at 3 university teaching hospitals where the participants were a convenience sample of internal medicine residents and fellows or attending physicians in general internal medicine or cardiology. For each scenario, physicians were asked to estimate the baseline 5-year risk of a coronary heart disease event and the revised risk if the patient were to receive lipid-lowering drug therapy. Estimates of the baseline 5-year risk were compared with values calculated from Framingham risk equations. Inaccurate responses were defined as those with a ratio of estimated to actual risk of more than 1.5 or less than 0.67. Physicians' estimates of the relative risk reduction with therapy were considered to be accurate if they were between 25% and 40%.

Results

79 physicians (53 residents, 8 fellows, 18 attending physicians) completed the survey. Only 24% of physicians' risk estimates were accurate. In most cases, physicians overestimated the absolute risk of cardiovascular events without therapy (proportion overestimating ranged from 32-92% for the 12 individual scenarios). Physicians made larger errors in patient scenarios involving patients with high total or LDL cholesterol levels. Physicians' estimates of the relative risk reduction from treatment were more accurate: 43% of estimates were between 25 and 40%. Over 85% of physicians recommended treatment in 10 of 12 scenarios.

Conclusions

Physicians overestimate the absolute risk of CHD events and the potential absolute benefit of drug therapy.

Background

Recent data suggest that aspirin may be effective for reducing cancer mortality.

Objective

To examine whether including a cancer mortality-reducing effect influences which men would benefit from aspirin for primary prevention.

Design

We modified our existing Markov model that examines the effects of aspirin among middle-aged men with no previous history of cardiovascular disease or diabetes. For our base case scenario of 45-year-old men, we examined costs and life-years for men taking aspirin for 10 years compared with men who were not taking aspirin over those 10 years; after 10 years, we equalized treatment and followed the cohort until death. We compared our results depending on whether or not we included a 22 % relative reduction in cancer mortality, based on a recent meta-analysis. We discounted costs and benefits at 3 % and employed a third party payer perspective.

Main measure

Cost per quality-adjusted life year (QALY) gained.

Key results

When no effect on cancer mortality was included, aspirin had a cost per QALY gained of $22,492 at 5 % 10-year coronary heart disease (CHD) risk; at 2.5 % risk or below, no treatment was favored. When we included a reduction in cancer mortality, aspirin became cost-effective for men at 2.5 % risk as well (cost per QALY, $43,342). Results were somewhat sensitive to utility of taking aspirin daily; risk of death after myocardial infarction; and effects of aspirin on stroke, myocardial infarction, and sudden death. However, aspirin remained cost-saving or cost-effective (< $50,000 per QALY) in probabilistic analyses (59 % with no cancer effect included; 96 % with cancer effect) for men at 5 % risk.

Conclusions

Including an effect of aspirin on cancer mortality influences the threshold for prescribing aspirin for primary prevention in men. If such an effect is real, many middle-aged men at low cardiovascular risk would become candidates for regular aspirin use.

Background

The coronary artery calcium (CAC) score is an independent predictor of coronary heart disease. We sought to combine information from the CAC score with information from conventional cardiac risk factors to produce post-test risk estimates, and to determine whether the score may add clinically useful information.

Methods

We measured the independent cross-sectional associations between conventional cardiac risk factors and the CAC score among asymptomatic persons referred for non-contrast electron beam computed tomography. Using the resulting multivariable models and published CAC score-specific relative risk estimates, we estimated post-test coronary heart disease risk in a number of different scenarios.

Results

Among 9341 asymptomatic study participants (age 35-88 years, 40% female), we found that conventional coronary heart disease risk factors including age, male sex, self-reported hypertension, diabetes and high cholesterol were independent predictors of the CAC score, and we used the resulting multivariable models for predicting post-test risk in a variety of scenarios. Our models predicted, for example, that a 60-year-old non-smoking non-diabetic women with hypertension and high cholesterol would have a 47% chance of having a CAC score of zero, reducing her 10-year risk estimate from 15% (per Framingham) to 6-9%; if her score were over 100, however (a 17% chance), her risk estimate would be markedly higher (25-51% in 10 years). In low risk scenarios, the CAC score is very likely to be zero or low, and unlikely to change management.

Conclusion

Combining information from the CAC score with information from conventional risk factors can change assessment of coronary heart disease risk to an extent that may be clinically important, especially when the pre-test 10-year risk estimate is intermediate. The attached spreadsheet makes these calculations easy.

Abstract Background Compared with non-Latinos, Latinos in the US have low rates of colorectal cancer (CRC) screening and low rates of knowledge regarding CRC screening tests and guidelines. Spanish speaking Latinos have particularly low CRC screening rates and screening knowledge. Our purpose was twofold: (1) to evaluate the effect of a computer-based, Spanish-language CRC screening decision aid on screening knowledge, intent to obtain screening, and screening self-efficacy in a community sample of Latinos with limited English proficiency (LEP); and (2) to survey these decision aid viewers at four months to determine their rates of CRC discussions with a health care provider as well as their rates of screening test completion. Methods We recruited 50-75 year old Latinos with LEP who were not current with CRC. Participants screening viewed a 14 minute multimedia decision aid that addresses CRC screening rationale, recommendations, and options. We conducted an uncontrolled (pre-post) study in which we assessed screening knowledge, self-efficacy, and intent at baseline and immediately after decision aid viewing. We also conducted a follow-up telephone survey of participants at four months to examine rates of patient-provider screening discussions and test completion. Results Among n = 80 participants, knowledge scores increased from 20% (before) to 72% (after) decision aid viewing (absolute difference [95%CI]: 52% [46, 59]). The proportion with high screening self-efficacy increased from 67% to 92% (25% [13, 37]); the proportion with high screening intent increased from 63% to 95% (32% [21, 44]). We reached 68 (85%) of 80 participants eligible for the follow-up survey. Of these 36 (53%) reported discussing screening with a provider and 13 (19%) completed a test. Conclusion Viewing a Spanish-language decision aid increased CRC screening knowledge, self-efficacy, and intent among Latinos with LEP. Decision aid viewing appeared to promote both CRC screening discussions with health care providers and test completion. The decision aid may be an effective tool for promoting CRC screening and reducing screening disparities in this population.

Background

The coronary artery calcium (CAC) score predicts coronary heart disease (CHD) events, but methods for interpreting the score in combination with conventional CHD risk factors have not been established.

Methods and results

We analyzed CAC scores and CHD risk factor measurements from 6757 black, Chinese, Hispanic, and white men and women aged 45 to 84 years in the Multi-Ethnic Study of Atherosclerosis (MESA). CAC was associated with age, sex, race/ethnicity, and all conventional CHD risk factors. Multivariable models using these factors predicted the presence of CAC (C statistic=0.789) and degree of elevation (16% of variation explained) and can be used to update a "pretest" CHD risk estimate, such as the 10-year Framingham Risk Score, that is based on an individual's conventional risk factors. In scenarios in which a high CAC score is expected, a moderately elevated CAC score of 50 is reassuring (eg, reducing risk from 10% to 6% in a healthy older white man), but when a low/zero CAC score is expected, even with identical pretest CHD risk, the same CAC score of 50 may be alarmingly high (eg, increasing risk from 10% to 20% in a middle-aged black woman with multiple risk factors). Both the magnitude and direction of the shift in risk varied markedly with pretest CHD risk and with the pattern of risk factors.

Conclusions

Knowledge of what CAC score to expect for an individual patient, based on their conventional risk factors, may help clinicians decide when to order a CAC test and how to interpret the results.

Benefit-targeted statin prescribing may be superior to risk-targeted statin prescribing (the current standard), but the impact and efficiency of this approach are unclear. We analyzed the National Health and Nutrition Examination Survey (NHANES) using an open-source model (the Prevention Impact and Efficiency Model) to compare targeting of statin therapy according to expected benefit (benefit-targeted) versus baseline risk (risk-targeted) in terms of projected population-level impact and efficiency. Impact was defined as relative % reduction in atherosclerotic cardiovascular disease in the US population for the given strategy compared to current statin treatment patterns; and efficiency as the number needed to treat over 10 years (NNT10, average and maximum) to prevent each atherosclerotic cardiovascular disease event. Benefit-targeted moderate-intensity statin therapy at a treatment threshold of 2.3% expected 10-year absolute risk reduction could produce a 5.7% impact (95% confidence interval, 4.8-6.7). This is approximately equivalent to the potential impact of risk-targeted therapy at a treatment threshold of 5% 10-year atherosclerotic cardiovascular disease risk (5.6% impact [4.7-6.6]). Whereas the estimated maximum NNT10 is much improved for benefit-targeted versus risk-targeted therapy at these equivalent-impact thresholds (43.5 vs 180), the average NNT10 is nearly equivalent (24.2 vs 24.6). Reaching 10% impact (half the Healthy People 2020 impact objective, loosely defined) is theoretically possible with benefit-targeted moderate-intensity statins of persons with expected absolute risk reduction >2.3% if we expand age eligibility and account for treatment of all persons with diabetes mellitus or with low-density lipoprotein >190 mg/dL (impact=12.4%; average NNT10=23.0). Benefit-based targeting of statin therapy provides modest gains in efficiency over risk-based prescribing and could theoretically help attain approximately half of the Healthy People 2020 impact goal with reasonable efficiency.

Background

The coronary artery calcium (CAC) score predicts future coronary heart disease (CHD) events and could be used to guide primary prevention interventions, but CAC measurement has costs and exposes patients to low-dose radiation.

Methods and results

We estimated the cost-effectiveness of measuring CAC and prescribing statin therapy based on the resulting score under a range of assumptions using an established model enhanced with CAC distribution and risk estimates from the Multi-Ethnic Study of Atherosclerosis. Ten years of statin treatment for 10,000 55-year-old women with high cholesterol (10-year CHD risk, 7.5%) was projected to prevent 32 myocardial infarctions, cause 70 cases of statin-induced myopathy, and add 1108 years to total life expectancy. Measuring CAC and targeting statin treatment to the 2500 women with CAC>0 would provide 45% of the benefit (+501 life-years), but CAC measurement would cost $2.25 million and cause 9 radiation-induced cancers. Treat all was preferable to CAC screening in this scenario and across a broad range of other scenarios (CHD risk, 2.5%-15%) when statin assumptions were favorable ($0.13 per pill and no quality of life penalty). When statin assumptions were less favorable ($1.00 per pill and disutility=0.00384), CAC screening with statin treatment for persons with CAC>0 was cost-effective (<$50 000 per quality-adjusted life-year) in this scenario, in 55-year-old men with CHD risk 7.5%, and in other intermediate risk scenarios (CHD risk, 5%-10%). Our results were critically sensitive to statin cost and disutility and relatively robust to other assumptions. Alternate CAC treatment thresholds (>100 or >300) were generally not cost-effective.

Conclusions

CAC testing in intermediate risk patients can be cost-effective but only if statins are costly or significantly affect quality of life.

Background

Decision support interventions (DESIs) provide a mechanism to translate comparative effectiveness research results into clinical care so that patients are able to make informed decisions. Patient decision support interventions for prostate-specific antigen (PSA) have been shown to promote informed decision making and reduce PSA testing in efficacy trials, but their impact in real world settings is not clear.

Objective

We performed an effectiveness trial of PSA decision support interventions in primary care.

Design

A randomized controlled trial of three distribution strategies was compared to a control.

Participants

Participants included 2,550 men eligible for PSA testing (76.6 % of the eligible population) and 2001 survey respondents (60.1 % survey response rate).

Interventions

The intervention groups were: 1) mailed the DESI in DVD format, 2) offered a shared medical appointment (SMA) to view the DESI with other men and discuss, and 3) both options.

Main measures

We measured PSA testing identified via electronic medical record at 12 months and DESI use by self-report 4 months after the intervention mailing.

Key results

We found no differences in PSA testing across the three distribution strategies over a year-long follow-up period: 21 %, 24 %, 22 % in the DESI, SMA, and combined group respectively, compared to 21 % in the control group (p = 0.51). Self-reported DESI use was low across all strategies at 4 months: 16 % in the mailed DESI group, 6 % in the SMA group, and 15 % in the combined group (p = < 0.0001).

Conclusions

Mailing PSA decision support interventions or inviting men to shared medical appointments unrelated to a primary care office visit do not appear to promote informed decision making, or change PSA testing behavior.