Risk Adjustment with Multivariate Techniques – State of New York

1. Which factors are supposedly related to CABG (coronary artery bypass graft surgery) mortality?

The New York State Department of Health publishes certain multivariable risk factors for patients undergoing CABG, recognizing that certain factors potentially contribute to mortality rates following the procedure (“Adult Cardiac”, 2013). The defined risk factors are based on established research within the field of epidemiology (Fleming, 2008, p. 142; Hannan, 2006). For CABG patients, risk factors include demographics, hemodynamic state, comorbidities, severity of the atherosclerotic process, ventricular function and previous open-heart operations (Fleming, 2008, p. 150). Each of these factors influences the patient’s recovery in some way.

According to the text, various demographics impact the patient’s level of risk. Older patients present a higher risk for surgery, as they have less physical strength to recover from the procedure (Fleming, 2008, p. 142). Gender also plays an important role, with women being at higher risk for mortality than men (Bukkapatnam, 2010, p. 339). Ethnicity is an important risk factor for operative mortality, with Hispanics, Asians and African Americans at the greatest risk (Yeo et al., 2007, p. 59). Other studies also suggest that ethnicity (combined with cultural, racial and socioeconomic factors) similarly present unique risks to patients as barriers to care (Fleming, 2008, p. 143). A patient’s existing clinical characteristics, including factors such as their hemodynamic state (where patients in shock present a far higher risk), affect the patient’s risk level (Fleming, 2008, p. 143, 150). The progression of coronary artery disease also impacts the patient’s risk, with the presence of certain conditions decreasing the patient’s odds of surviving the procedure (Fleming, 2008, p. 150). Lastly, the patient’s previous medical history is significant, as patients with impaired ventricular function, or previous open-heart surgeries, present a different surgical risk (Fleming, 2008, p. 150). Overall, there are a variety of factors that should be taken into consideration when evaluating a patient’s mortality risk.

2. Which factors are most strongly related to CABG mortality?

Based on the Multivariate Approach to Risk Adjustment published by the New York State Department of Health, certain factors are most strongly related to CABG mortality. As reflected in the logistic regression, the single greatest risk factor for patients is hepatic failure (Fleming, 2008, p. 150). Patients with this condition are 21.190 times more likely than others to die during, or following, CABG (Fleming, 2008, p. 150). Shock is the second strongest factor related to CABG mortality. Patients in shock are 6.333 times more likely to die than those not in shock (Fleming, 2008, p. 150). Lastly, renal failure requiring dialysis is the third strongest factor related to CABG mortality (Fleming, 2008, p. 150). As reflected in the multivariate model, patients requiring dialysis are 5.688 more likely to die from CABG (Fleming, 2008, p. 150). Interestingly, for each of the strongest factors, it is either the presence or complete absence of the risk factor that contributes to the mortality rate (“Cardiac”, 2013, p. 12). There is no middle ground for these factors.

3. How could one derive an expected mortality rate from the multivariate model?

As explained by O’Connor et al. (1992) the statistical information in the multivariate model can be used to calculate an expected mortality rate for a patient as follows:

Odds = exp {0.793 + (0.0671 x age [years]) + (0.5105 x gender) + (insert coefficient x hemodynamic state) + (insert coefficient x comorbidity score) + (insert coefficient x severity) + (insert coefficient x ventricular function) + (0.6800 x prior CABG)} (p. 2113).

References

Adult Cardiac Surgery in New York State 2008-2010. (2013, August). New York State Department of Health. Retrieved from http://www.health.ny.gov/statistics/diseases/cardiovascular/heart_disease/docs/2008-2010_adult_cardiac_surgery.pdf

Bukkapatnam, R. N., Yeo, K. K., Li, Z., & Amsterdam, E. A. (2010). Operative mortality in women and men undergoing coronary artery bypass grafting (from the California coronary artery bypass grafting outcomes reporting program). The American Journal of Cardiology, 105(3), 339-342. Retrieved from http://dx.doi.org/10.1016/j.amjcard.2009.09.035

Fleming, S. T. (2008). Managerial epidemiology concepts and cases (2nd ed.). Chicago: Health Administration Press.

Hannan, E. L., Smith, C. R., Isom, O. W., Higgins, R. S., Gold, J. P., Culliford, A. T., et al. (2006). Risk stratification of in-hospital mortality for coronary artery bypass graft surgery. Journal of the American College of Cardiology, 47(3), 661-668.

O'Connor, G. T., Tryzelaar, J. F., Levy, D. G., Nowicki, E. R., Maloney, C. T., Morton, J. R., et al. (1992). Multivariate prediction of in-hospital mortality associated with coronary artery bypass graft surgery. Northern New England Cardiovascular Disease Study Group. Circulation, 85(6), 2110-2118.

Yeo, K. K., Li, Z., & Amsterdam, E. (2007). Clinical characteristics and 30-day mortality among Caucasians, Hispanics, Asians, and African Americans in the 2003 California Coronary Artery Bypass Graft Surgery Outcomes Reporting Program. The American Journal of Cardiology, 100(1), 59-63. Retrieved from http://dx.doi.org/10.1016/j.amjcard.2007.02.053