Impact of risk scoring on decision-making in symptomatic moderate carotid atherosclerosis

Decision analytic modeling was useful to assess the impact of a prediction model on health outcomes before a randomized trial

OBJECTIVE

To demonstrate how decision analytic models (DAMs) can be used to quantify impact of using a (diagnostic or prognostic) prediction model in clinical practice and provide general guidance on how to perform such assessments.

STUDY DESIGN AND SETTING

A DAM was developed to assess the impact of using the HEART score for predicting major adverse cardiac events (MACE). Impact on patient health outcomes and health care costs was assessed in scenarios by varying compliance with and informed deviation (ID) (using additional clinical knowledge) from HEART score management recommendations. Probabilistic sensitivity analysis was used to assess estimated impact robustness.

RESULTS

Impact of using the HEART score on health outcomes and health care costs was influenced by an interplay of compliance with and ID from HEART score management recommendations. Compliance of 50% (with 0% ID) resulted in increased missed MACE and costs compared with usual care. Any compliance combined with at least 50% ID reduced both costs and missed MACE. Other scenarios yielded a reduction in missed MACE at higher costs.

CONCLUSION

Decision analytic modeling is a useful approach to assess impact of using a prediction model in practice on health outcomes and health care costs. This approach is recommended before conducting an impact trial to improve its design and conduct.

Associations Among Practice Variation, Clinician Characteristics, and Care Algorithm Usage: A Multispecialty Vignette Study

The objective was to quantitatively evaluate clinician characteristics associated with unwarranted practice variation, and how clinical care algorithms influence this variation. Participants (142 physicians, 53 nurse practitioners, and 9 physician assistants in family medicine, internal medicine, and cardiology) described their management of 4 clinical vignettes, first based on their own practice (unguided), then using care algorithms (guided). The authors quantitatively estimated variation in management. Cardiologists demonstrated 17% lower variation in unguided responses than generalists (fold-change 0.83 [95% confidence interval (CI) 0.68, 0.97]), and those who agreed that practice variation can realistically be reduced had 16% lower variation than those who did not (fold-change 0.84 [CI, 0.71, 0.99]). A 17% reduction in variation was observed for guided responses compared with baseline (unguided) responses (fold-change 0.83 [CI, 0.76, 0.90]). Differences were otherwise similar across clinician subgroups and attitudes. Unwarranted practice variation was similar across most clinician subgroups. The authors conclude that care algorithms can reduce variation in management.

An Accumulated Deficits Model Predicts Perioperative and Long-term Adverse Events after Carotid Endarterectomy

BACKGROUND

There is increasing recognition that decreased reserve in multiple organ systems, known as accumulated deficits (AD), may better stratify perioperative risk than traditional risk indices. We hypothesized that an AD model would predict both perioperative adverse events and long-term survival after carotid endarterectomy (CEA), particularly important in asymptomatic patients.

METHODS

Consecutive patients undergoing CEA between 1st January 2000 and 31st December 2010 were retrospectively identified. Seven of the deficit items from the Canadian Study of Health and Aging-frailty index (coronary disease, renal insufficiency, pulmonary disease, peripheral vascular disease, heart failure, hypertension, and diabetes) were tabulated for each patient. Predictors of perioperative and long-term outcomes were evaluated using regression analysis.

RESULTS

About 1,782 CEAs in 1,496 patients (mean age: 71.3 ± 9.3 years, 56.3% male, 35.4% symptomatic) were included. The risk of major adverse events (stroke, death, or myocardial infarction) at 30 days for patients with ≤3 deficits was 2.53% vs. 8.81% for patients with ≥4 deficits (P < 0.001). For patients with ≥5 deficits, the risk was 15.18%. Each additional deficit increased the odds of a 30-day major adverse event and hospital stay >2 days by 1.64 (P < 0.001) and 1.15 (P < 0.001), respectively. In multivariate analysis, the presence of ≥4 deficits was more predictive of perioperative major adverse events (odds ratio [OR] = 3.62, P < 0.001) than symptomatology within 6 months (OR = 1.57, P = 0.08) or octogenarian status (OR = 2.00, P = 0.02). Kaplan-Meier analysis showed significantly decreased survival over time with accumulating deficits (P < 0.001). Patients with ≥4 deficits have a hazards ratio for death of 2.6 compared to patients with ≤3 deficits (P < 0.001). Overall survival is estimated at 79.5% (95% confidence interval [CI]: 0.77-0.82) at 5 years in patients with ≤3 deficits versus 52.4% (95% CI: 0.46-0.58) in patients with ≥4 deficits, respectively. In subgroup analysis of asymptomatic patients, 5-year survival for octogenarian male patients with ≥4 deficits was only 26.8%. For asymptomatic males aged 70-79 years with ≥4 deficits, 5-year survival was 59.9%.

CONCLUSIONS

An AD model is more predictive of perioperative adverse events after CEA than age or symptomatic status. This model remains predictive of long-term survival. In asymptomatic male octogenarians with 4 or more AD, 5-year survival is severely limited.