Background
Mindstep is an app that aims to improve dementia screening by assessing cognition and risk factors. It considers important clinical risk factors, including prodromal symptoms, mental health disorders, and differential diagnoses of dementia. The 9-item Patient Health Questionnaire for depression (PHQ-9) and the 7-item Generalized Anxiety Disorder Scale (GAD-7) are widely validated and commonly used scales used in screening for depression and anxiety disorders, respectively. Shortened versions of both (PHQ-2/GAD-2) have been produced.
Objective
We sought to develop a method that maintained the brevity of these shorter questionnaires while maintaining the better precision of the original questionnaires.
Methods
Single questions were designed to encompass symptoms covered in the original questionnaires. Answers to these questions were combined with PHQ-2/GAD-2, and anonymized risk factors were collected by Mindset4Dementia from 2235 users. Machine learning models were trained to use these single questions in combination with data already collected by the app: age, response to a joke, and reporting of functional impairment to predict binary and continuous outcomes as measured using PHQ-9/GAD-7. Our model was developed with a training data set by using 10-fold cross-validation and a holdout testing data set and compared to results from using the shorter questionnaires (PHQ-2/GAD-2) alone to benchmark performance.
Results
We were able to achieve superior performance in predicting PHQ-9/GAD-7 screening cutoffs compared to PHQ-2 (difference in area under the curve 0.04, 95% CI 0.00-0.08, P=.02) but not GAD-2 (difference in area under the curve 0.00, 95% CI –0.02 to 0.03, P=.42). Regression models were able to accurately predict total questionnaire scores in PHQ-9 (R2=0.655, mean absolute error=2.267) and GAD-7 (R2=0.837, mean absolute error=1.780).
Conclusions
We app-adapted PHQ-4 by adding brief summary questions about factors normally covered in the longer questionnaires. We additionally trained machine learning models that used the wide range of additional information already collected in Mindstep to make a short app-based screening tool for affective disorders, which appears to have superior or equivalent performance to well-established methods.
Funding Acknowledgements
Type of funding sources: Other. Main funding source(s): BRAVO trial: BHF SP/10/002/28189, FS/10/038, FS/11/92/29122, FS/13/44/30291) National Institute for Health Research Imperial Biomedical Research Centre. HOPE-HF trial: British Heart Foundation (CS/15/3/31405, FS/13/44/30291, FS/15/53/31615, FS/14/27/30752, FS/10/038).
Introduction
The optimal atrioventricular (AV) delay for implantable cardiac devices can be derived by echocardiography or beat-by-beat blood pressure measurements. However, both of these approaches are labour intensive and neither could be incorporated into an implantable cardiac device for frequent repeated optimisations. Laser Doppler perfusion monitoring (LDPM) measures blood flow through tissue. LDPM has been miniaturised ready to be incorporated into future implantable cardiac devices.
Purpose
We studied if LDPM is a clinically reliable alternative method to blood-pressure measurements to determine optimal AV delay.
Methods
Data from 58 patients undergoing 94 clinical AVD optimisations using LDPM and simultaneous non-invasive beat-by-beat blood pressure was obtained. The optimal AV delay for each method and for each optimisation was determined using a curve of haemodynamic response to switching from AAI (reference state) to DDD (test state) at a series of AV delays (40, 80, 120, 160, 200, 240 ms). We then compared the derived optimal AV delays between the two measurement approaches. We also assessed the impact of the paced heart-rate on agreement between laser Doppler and Blood-Pressure derived optimal AV delays.
Results
The AV delay derived using LDPM was not clinically significant different from that derived by blood pressure changes. The median difference was -9ms (IQR -26 to 7, p = 0.05). Variability between the two methods was low (median absolute deviation 17ms). Optimisations performed at higher heart-rates resulted in a non-significant smaller difference between the LDPM and blood-pressure derived AV delays (median absolute deviation 12 vs 22 ms, p = 0.11).
Conclusions
Optimal AVDs derived from non-invasive blood-pressure or laser Doppler perfusion methods are clinically equivalent. The addition of laser Doppler to future implantable cardiac devices may enable devices to dynamically and reliably optimise AV delays. Abstract Figure 1
BACKGROUND
Mindset4Dementia is an app that aims to improve dementia screening by assessing cognition and risk factors. It considers important clinical risk factors, including prodromal symptoms, mental health disorders, and differential diagnoses of dementia. The PHQ-9 and GAD-7 are widely validated, and commonly used scales used in screening for depression and anxiety disorders respectively. Shortened versions of both (PHQ-2/GAD-2) have been produced.
OBJECTIVE
We sought to develop a method that maintained the brevity of these shorter questionnaires while maintaining the better precision of the original questionnaires
METHODS
Single questions were designed to encompass symptoms covered in the original questionnaires. Answers to these questions were combined with the PHQ-2/GAD-2 and anonymized risk factors collected by Mindset4Dementia. Machine learning models were trained to use these single questions in combination with data already collected by the app - age, response to a joke and reporting of functional impairment to predict binary and continuous outcomes as measured by the PHQ-9/GAD-7. Our model was developed with a training dataset using ten-fold cross-validation and a hold-out testing datasets and compared to results from using the shorter questionnaires (PHQ-2/GAD-2) alone to benchmark performance.
RESULTS
We were able to achieve superior performance in predicting PHQ-9/GAD-7 screening cut-offs than the PHQ-2 (difference In AUC 0.04, 95% CI 0.00 – 0.08, P = 0.02) but not to GAD-2 (difference in AUC 0.00, 95% CI -0.02 – 0.03, P = 0.42). In regression models we were able to accurately predict total questionnaire scores; PHQ-9 (R2 = 0.655, MAE = 2.267), GAD-7 (R2 = 0.837, MAE = 1.780).
CONCLUSIONS
We have developed a short screening tool for affective disorders with superior or equivalent performance to well established methods.
scite is a Brooklyn-based organization that helps researchers better discover and understand research articles through Smart Citations–citations that display the context of the citation and describe whether the article provides supporting or contrasting evidence. scite is used by students and researchers from around the world and is funded in part by the National Science Foundation and the National Institute on Drug Abuse of the National Institutes of Health.