Continuous updating of individual headache forecasting models using Bayesian methods

Published on Sep 1, 2021in Headache4.041
· DOI :10.1111/HEAD.14182
Timothy T. Houle40
Estimated H-index: 40
(Harvard University),
Hao Deng12
Estimated H-index: 12
(Harvard University)
+ 1 AuthorsDana P. Turner13
Estimated H-index: 13
(Harvard University)
Sources
Abstract
OBJECTIVE To illustrate the benefits of deploying individual headache forecasting models using continuous updating with Bayesian methods. BACKGROUND The ability to reliably forecast headache attacks within an individual over time would enhance the study of attacks and allow preemptive treatment. However, deploying a suitable forecasting model in a clinical setting will likely involve several unique challenges related to heterogeneity in the predictor weights, limited or sparse data, and the need for a quick "warm-up." The use of Bayesian methods offers solutions to each of these specific challenges. METHODS This was a post hoc analysis of a cohort study of individuals with episodic migraine attacks. Individuals completed daily diaries that allowed the estimation of several forecasting models, each using different types of ancillary information incorporated into formal prior probability distributions. An in silico analysis was conducted that mimicked the deployment of these models in a clinical-like setting where the parameters of the models were continuously updated and evaluated each day using root mean square error (RMSE). RESULTS Individuals (N = 95) were followed for 50 days and contributed 3359 days of nonmissing diary data. During the observation period, there were 1293/3359 (38.5%) days with a headache attack. Self-reported baseline headache frequency was associated with the corresponding predicted probability of future attacks, r = 0.15-0.39. At Day 25, the correlation between baseline information and predicted attack likelihood was r = 0.29 (95% CI: 0.09-0.47). Additionally, the use of prior probability distributions for model parameters improved the model fit, especially early in the deployment of the models (e.g., Day 5 RMSE 0.45 vs. 0.43). Models using informative prior probability distributions outperformed the models estimated without this information during the first 42 days of observation, although performance became more similar as more data were collected. CONCLUSIONS This analysis demonstrates the value of Bayesian methods in using additional available information to improve forecasting model performance, especially early in the deployment of a forecasting model. To obtain the full value of such models or to apply any model in clinical settings, a model with adequate discrimination and calibration will be needed.
📖 Papers frequently viewed together
6 Citations
1 Citations
References24
Newest
#1Dana P. TurnerH-Index: 13
#2Hao Deng (Johns Hopkins University)H-Index: 12
Last. Timothy T. HouleH-Index: 40
view all 3 authors...
3 CitationsSource
#1Dana P. Turner (Harvard University)H-Index: 13
#2Lisa Leffert (Harvard University)H-Index: 27
Last. Timothy T. Houle (Harvard University)H-Index: 40
view all 3 authors...
2 CitationsSource
#1Dana P. Turner (Harvard University)H-Index: 13
#2Adriana D. Lebowitz (Harvard University)H-Index: 2
Last. Timothy T. Houle (Harvard University)H-Index: 40
view all 4 authors...
OBJECTIVES: To examine the hypothesis that surprising experiences of headache triggers are associated with daily headache activity. BACKGROUND: Little is known about the specific environmental or behavioral interactions that might trigger a headache attack in a prone individual. We propose that headache trigger exposures can be usefully characterized, not only by their mechanisms of action, but also on the degree of "surprise" they present to the individual. This hypothesis is based on elements ...
13 CitationsSource
#1Elizabeth K. Seng (Yeshiva University)H-Index: 18
#2Pablo PrietoH-Index: 1
Last. Marina Vives-MestresH-Index: 2
view all 4 authors...
OBJECTIVE: To evaluate factors associated naturalistically with adherence to a mobile headache diary. BACKGROUND: Self-monitoring (keeping a headache diary) is commonly used in headache to enhance diagnostic accuracy and evaluate the effectiveness of headache therapies. Mobile applications are increasingly used to facilitate keeping a headache diary. Little is known about the factors associated with adherence to mobile headache diaries. METHODS: In this naturalistic longitudinal cohort study, pe...
16 CitationsSource
#1Dana P. Turner (Harvard University)H-Index: 13
#2Adriana D. Lebowitz (Harvard University)H-Index: 2
Last. Timothy T. Houle (Harvard University)H-Index: 40
view all 4 authors...
Purpose of Review This review synthesizes the utility of measuring migraine triggers for the purpose of forecasting future headache attacks. The nature of forecasting models, headache triggers as inputs to such models, and how these trigger exposures can be measured for forecasting are reviewed. A critical evaluation of the existing forecasting models in the context of their potential application for preemptive treatment is considered.
10 CitationsSource
#1Daniel SerranoH-Index: 27
#2Richard B. Lipton (Albert Einstein College of Medicine)H-Index: 178
Last. Dawn C. Buse (Albert Einstein College of Medicine)H-Index: 53
view all 7 authors...
Relatively little is known about the stability of a diagnosis of episodic migraine (EM) or chronic migraine (CM) over time. This study examines natural fluctuations in self-reported headache frequency as well as the stability and variation in migraine type among individuals meeting criteria for EM and CM at baseline. The Chronic Migraine Epidemiology and Outcomes (CaMEO) Study was a longitudinal survey of US adults with EM and CM identified by a web-questionnaire. A validated questionnaire was u...
91 CitationsSource
#1Leonhard Held (UZH: University of Zurich)H-Index: 46
#2Sebastian Meyer (FAU: University of Erlangen-Nuremberg)H-Index: 16
Last. Johannes Bracher (UZH: University of Zurich)H-Index: 6
view all 3 authors...
Routine surveillance of notifiable infectious diseases gives rise to daily or weekly counts of reported cases stratified by region and age group. From a public health perspective, forecasts of infectious disease spread are of central importance. We argue that such forecasts need to properly incorporate the attached uncertainty, so they should be probabilistic in nature. However, forecasts also need to take into account temporal dependencies inherent to communicable diseases, spatial dynamics thr...
52 CitationsSource
#1Richard B. Lipton (Albert Einstein College of Medicine)H-Index: 178
#2Jelena M. Pavlovic (Albert Einstein College of Medicine)H-Index: 17
Last. Dawn C. Buse (Albert Einstein College of Medicine)H-Index: 53
view all 3 authors...
6 CitationsSource
#1Timothy T. HouleH-Index: 40
#2Dana P. TurnerH-Index: 13
Last. Charles H. TegelerH-Index: 28
view all 7 authors...
Objective To develop and validate a prediction model that forecasts future migraine attacks for an individual headache sufferer. Background Many headache patients and physicians believe that precipitants of headache can be identified and avoided or managed to reduce the frequency of headache attacks. Of the numerous candidate triggers, perceived stress has received considerable attention for its association with the onset of headache in episodic and chronic headache sufferers. However, no eviden...
35 CitationsSource
#1Sander Greenland (UCLA: University of California, Los Angeles)H-Index: 130
#2Mohammad Ali Mansournia (Tehran University of Medical Sciences)H-Index: 38
Last. Douglas G. Altman (University of Oxford)H-Index: 268
view all 3 authors...
Effects of treatment or other exposure on outcome events are commonly measured by ratios of risks, rates, or odds. Adjusted versions of these measures are usually estimated by maximum likelihood regression (eg, logistic, Poisson, or Cox modelling). But resulting estimates of effect measures can have serious bias when the data lack adequate case numbers for some combination of exposure and outcome levels. This bias can occur even in quite large datasets and is hence often termed sparse data bias....
403 CitationsSource
Cited By0
Newest