|Category Reference for Presentations|
|AHE||Applied Health Economics||DEC||Decision Psychology and Shared Decision Making|
|HSP||Health Services, and Policy Research||MET||Quantitative Methods and Theoretical Developments|
* Candidate for the Lee B. Lusted Student Prize Competition
Purpose: Because of the non-linear relationship between some model inputs and outcomes, previous research using numerical methods suggested that the choice of modeling technique — cohort versus individual — can have significant effects on model results. However, the direction of the effect is not known a priori. Our purpose was to derive mathematically (analytically) the conditions under which a cohort model (which does not capture baseline heterogeneity) compared with an individual-based simulation approach will under-estimate or over-estimate health outcomes.
Method: We used a three-state Markov model to estimate the cost-effectiveness of a hypothetical intervention, with efficacy e and cost I, to prevent disease developing at rate b resulting in disease-specific death at rate d, cost at c per period, and quality of life loss q. All parameters, including all-cause mortality at rate m, were constant. We solved the continuous-time model analytically and derived expressions for life expectancy, discounted quality-adjusted life years (QALYs), discounted lifetime disease costs, incremental cost-effectiveness ratio (ICER), and net monetary benefits (NMB). An outcome was calculated using the mean of the input under the cohort-based approach and the whole input distribution for all persons under the individual-based approach. We investigated the impact of heterogeneity on outcomes by varying one parameter at a time (e.g., b) while keeping all others constant (e.g., e, m, d, q, and c). We evaluated the curvature of outcome functions and used Jensen's inequality (i.e. if f is convex in X, f(E[X] £ E[f(X)]) to determine whether a cohort model under- or over-estimated a heath outcome (e.g., if f was convex, then a cohort model under-estimated the outcome).
Both life expectancy and QALYs were underestimated by the cohort-based approach (Figure1). If there was only heterogeneity in disease hazard, discounted costs were overestimated whereas QALYs gained, incremental costs, and ICER were under or overestimated depending on the value of b. ICER was overestimated and NMB underestimated when there was heterogeneity in efficacy only. Both approaches yielded the same outcome when there was only heterogeneity in c or q.
Conclusion: Use of a cohort-based approach that does not adjust for heterogeneity underestimates life expectancy and may under- or over-estimate other outcomes. Characterizing the bias is useful for calibration and predictions.
Purpose: Coronary heart disease (CHD) is the leading cause of morbidity and mortality in the United States and worldwide. While public health programs have substantially reduced CHD mortality rates since the 1970s, recent policy changes promoting accountable care organizations (ACOs) and integrated healthcare systems afford new opportunities to potentially improve population health through clinic- or hospital-based point-of-care interventions. To quantify the potential impact of these programs on population health, we identified gaps in preventive practices for CHD and projected health outcomes associated with adoption of point-of-care interventions addressing these gaps.
Method: We analyzed national data from 3.3 million patients who completed Life Line Screening health questionnaires between 2003-2008. We examined the prevalence of unhealthy behaviors (smoking, sedentary lifestyle, poor diet) and clinical risk factors (systolic blood pressure>140mmHg, overweight or obese BMI, absence of statin use in CHD) and projected four-year risk of onset or progression of CHD (defined as angina, myocardial infarction, or cardiac death) using Framingham Heart Study models and meta-analyses of randomized trials. We linked participants' residential addresses to U.S. Census data at the zip code-level to adjust for community income, education, and racial/ethnic diversity.
Result: Median age was 64 years (interquartile range, 56-71), 64% were women, 10% had diabetes, 25% smoked, and 9% had a history of CHD or stroke. Full adherence to preventive behaviors and practices was present in <2% of patients. Using ordinal logistic regression, we found that poorer adherence was associated with age>65 years (aOR 1.3, 95% CI 1.2-1.3), male gender (aOR 1.3, 1.2-1.3), black race (aOR 1.3, 1.3-1.4), and diabetes (aOR 1.4, 1.4-1.5). Adjusting for patient compliance and uptake, clinic- or hospital-based distribution of free high-dose nicotine patches to smokers, free statins to patients with CHD, free pedometers to all patients, and promotion of a Mediterranean diet high in nuts would prevent at least 92, 110, 490, and 480 new or recurrent incidents of CHD per-100,000 adults over four years. In comparison, a 50% reduction in salt consumption--a recommendation widely endorsed by public health organizations--would prevent at least 380 incidents.
Conclusion: Adherence to recommendations for the prevention or progression of CHD is poor. In light of growing integration of healthcare systems and ACOs, selective adoption of clinic- or hospital-based point-of-care programs may further reduce the population burden of CHD.
Purpose: Comparing surgical performance across institutions may assist in identifying best practices in the interest of patients, clinicians, payers and policy makers. Although 30-day mortality is the most commonly used performance indicator, follow-up length required to capture deaths potentially due to surgical quality is debated. We aimed to determine optimal follow-up period for mortality after esophageal resection to measure surgical quality, using a decision analytical approach. We also aimed to develop a case-mix adjustment model to compare surgical performance across hospitals.
Methods: Esophageal cancer patients were included prospectively from a large university hospital. The cause of death was classified for all patients who died within one year after surgery as potentially surgery related vs. probably not surgery related (progression of disease). We calculated the numbers of false positives and false negatives for a range of follow-up periods and the ratio between these. Case-mix adjustment models were developed using logistic regression.
Results: Between 1991 and 2011 1,283 patients were operated, of whom 393 (29%) died within 1 year: 107 (29%) with a potential surgical cause, and 261 (70%) with a most likely oncological cause. For 5 deaths (1%), the cause was unknown. Within 30 days after surgery, 36 (2.9%) patients had died, all -except one unknown- with a surgical cause. Within 90 days after surgery 90 (7%) patients had died, 79 with a surgical cause and 10 with an oncological cause. With a follow-up period of 30 days, there would be 34/107 (33%) true positives (surgical deaths) against no false positives (oncological deaths), implying a ratio of infinity for the importance of misclassifying a death as potentially surgery related (figure). With 90 days follow-up, there would be 76 /107 (71%) true positives, and 8/261 (3%) false negatives, implying a ratio of 9.5. A ratio of 1:1 was only reached after 200 days. Important predictors of 90-day mortality were older age, male gender, involvement of resection margins, transthoracic surgery (vs. transhiatal surgery), weight loss prior to surgery, neoadjuvant therapy, and a history of cardiovascular disease.
Conclusions: A substantial number of surgery related deaths occurred after 30 days after surgery. The optimal length of follow-up depends on the misclassification costs attached to false-positives vs. false-negatives. Weighting both as equally important would require a follow-up up to 200 days.
Adalimumab and infliximab have been approved for the treatment of ulcerative colitis by several regulatory bodies across the globe. Both are viable treatment options when patients become refractory or intolerant to conventional therapy. Since placebo controlled randomized clinical trials for adalimumab were only published recently, the comparative effectiveness of adalimumab versus infliximab has not yet been established via accepted meta-analytic indirect comparison methods.
An exhaustive search strategy covered major medical databases and recent conventional meta-analyses to identify eligible randomized clinical trials (RCTs). A Bayesian random-effects indirect comparison meta-analysis was performed for five selected and patient-important clinical outcomes at 8 weeks and 52 weeks. Odds ratio (OR) estimates and associated 95% credible intervals (CrI) were produced.
Five eligible RCTs were identified from which data on clinical remission, clinical response, mucosal healing, response on the inflammatory bowel disease questionnaire (IBDQ), colectomy, serious adverse events, and discontinuation due to adverse events were extracted at 8 weeks and 52 weeks. For all efficacy outcomes at both time points, the indirect comparison meta-analysis of adalimumab versus infliximab favoured infliximab. At 8 weeks, clinical remission (OR=0.42, 95% CrI 0.17-0.97), clinical response (OR=0.45, 95% CrI 0.23-0.89) and mucosal healing (OR=0.46, 95% CrI 0.25-0.86) were all statistically significant, whereas IBDQ was not. At 52 weeks, OR estimates for all efficacy outcomes favoured infliximab, but since only two trials provided data for this time point, the results were not statistically significant. Sustained remission was also more likely with infliximab, and sustained response was statistically significantly more likely with infliximab (OR=0.53, 95% CrI 0.24-0.98). For serious adverse events and discontinuations due to adverse events, adalimumab and infliximab both trended towards smaller risk than placebo, but the findings were not statistically significant. Further, the indirect comparison of the adalimumab and infliximab yielded odds ratios close to 1.00 with wide credible intervals.
Our findings suggest that infliximab is significantly more effective than adalimumab for the treatment of moderately to severe ulcerative colitis at 8 weeks, is numerically more effective at 52 weeks, and is more effective at sustaining 8 weeks outcomes till 52 weeks. Findings on safety outcomes were statistically inconclusive.
Purpose: We investigated how using filtered longitudinal data as input for logistic regression to predict glaucoma progression affects the classification ability of the logistic regression function.
Methods: A Kalman filter was developed to reduce the process and measurement noise present in longitudinal data from the Collaborative Initial Glaucoma Treatment Study (CIGTS), a randomized clinical trial of patients with early to moderate open angle glaucoma (OAG). These filtered repeated measures estimates were then used as data input for logistic regression via generalized estimating equations in order to predict OAG progression in patients. Analysis of the receiver operating characteristic (ROC) curve was used to compare this Kalman filter-based model against the standard methodology of using raw observations from the clinical trial as data input for logistic regression.
Results: The Kalman filter-based model resulted in higher specificity and sensitivity compared to the standard raw observations logistic regression model. The area under the ROC curve (AUC) for the Kalman filter-based model was 0.953 while the AUC for the raw observations model was 0.890.
Conclusions: Kalman filtering for estimating the true value of disease-related variables has been shown to improve the progression identification ability of logistic regression functions as compared to the standard approach of using raw data. This approach is applicable to any chronic disease which is subject to noisy observations and requires longitudinal follow-up for effective disease management.
Purpose: Fresh embryo transfer (ET) is currently the norm for in vitro fertilization (IVF). A growing body of literature, however, suggests that pregnancies resulting from frozen-thawed ET have better obstetric and perinatal outcomes than pregnancies resulting from fresh ET. Better outcomes may be related to improved endometrial receptivity and placentation in the more physiologic hormonal environment associated with frozen ET. We used decision analysis to assess the effect of performing elective frozen ET for all IVF cycles.
Methods: A decision-analytic model was developed to compare complication rates for fresh and frozen ET in hypothetical cohorts of women of differing ages (<35, 35-37, 38-40, 41-42, and 43-44). Live-birth rates were derived from 2010 U.S. CDC data. Rates of preterm birth, preeclampsia, and small for gestational age infants based on maternal age and type of IVF cycle were derived from the literature. One-way and multi-way sensitivity analyses were performed to test the robustness of our findings.
Results: For women under 35, fresh and frozen ET result in a similar number of uncomplicated live births. For women aged 35-37, 38-40, 41-42, and 43-44, frozen ET was associated with a higher rate of uncomplicated live births (table). These results were sensitive to the protective effect of frozen ET on the likelihood of complications as well as live-birth rate per cycle with fresh and frozen ET. If the total live-birth rate with fresh ET was more than 3% (35-37), 2.6% (38-40), 1.6% (41-42), or 0.7% (43-44) higher than the total live-birth rate with frozen ET, fresh ET became the preferred strategy for maximizing likelihood of an uncomplicated live birth.
Conclusions: Frozen-thawed embryo transfer optimizes the uncomplicated live-birth rate in women over 35 undergoing IVF. These results are sensitive to the relative likelihood of live birth with fresh versus frozen ET and the protective effect of frozen ET on perinatal complications. Along with individual embryo quality, this data should be considered when contemplating the optimal IVF strategy for women 35 and older. In women under 35, the higher live-birth rate of fresh ET should be weighed against the higher risk of perinatal complications. Additional prospective research is needed to assess live-birth and complication rates following elective frozen-thawed embryo transfer.