Harnessing Consumer Wearable Digital Biomarkers for Individualized Recognition of Postpartum Depression Using the All of Us Research Program Data Set: Cross-Sectional Study

Abstract

Background: Postpartum depression (PPD) poses a significant maternal health challenge. The current approach to detecting PPD relies on in-person postpartum visits, which contributes to underdiagnosis. Furthermore, recognizing PPD symptoms can be challenging. Therefore, we explored the potential of using digital biomarkers from consumer wearables for PPD recognition.

Objective: The main goal of this study was to showcase the viability of using machine learning (ML) and digital biomarkers related to heart rate, physical activity, and energy expenditure derived from consumer-grade wearables for the recognition of PPD.

Methods: Using the All of Us Research Program Registered Tier v6 data set, we performed computational phenotyping of women with and without PPD following childbirth. Intraindividual ML models were developed using digital biomarkers from Fitbit to discern between prepregnancy, pregnancy, postpartum without depression, and postpartum with depression (ie, PPD diagnosis) periods. Models were built using generalized linear models, random forest, support vector machine, and k-nearest neighbor algorithms and evaluated using the κ statistic and multiclass area under the receiver operating characteristic curve (mAUC) to determine the algorithm with the best performance. The specificity of our individualized ML approach was confirmed in a cohort of women who gave birth and did not experience PPD. Moreover, we assessed the impact of a previous history of depression on model performance. We determined the variable importance for predicting the PPD period using Shapley additive explanations and confirmed the results using a permutation approach. Finally, we compared our individualized ML methodology against a traditional cohort-based ML model for PPD recognition and compared model performance using sensitivity, specificity, precision, recall, and F₁-score.

Results: Patient cohorts of women with valid Fitbit data who gave birth included <20 with PPD and 39 without PPD. Our results demonstrated that intraindividual models using digital biomarkers discerned among prepregnancy, pregnancy, postpartum without depression, and postpartum with depression (ie, PPD diagnosis) periods, with random forest (mAUC=0.85; κ=0.80) models outperforming generalized linear models (mAUC=0.82; κ=0.74), support vector machine (mAUC=0.75; κ=0.72), and k-nearest neighbor (mAUC=0.74; κ=0.62). Model performance decreased in women without PPD, illustrating the method's specificity. Previous depression history did not impact the efficacy of the model for PPD recognition. Moreover, we found that the most predictive biomarker of PPD was calories burned during the basal metabolic rate. Finally, individualized models surpassed the performance of a conventional cohort-based model for PPD detection.

Conclusions: This research establishes consumer wearables as a promising tool for PPD identification and highlights personalized ML approaches, which could transform early disease detection strategies.

Keywords: All of Us; Fitbit; machine learning; mobile phone; postpartum depression; wearable device.

Figure 1

An overview of the analysis workflow to evaluate the potential for digital biomarkers in postpartum depression (PPD) recognition. (1) Develop and perform computational phenotyping of PPD and non-PPD cohorts; (2) merge with available digital biomarker data for each woman (heart rate, steps, physical activity, and calories burned); (3) classify each day as 1 of 4 periods (prepregnancy period, pregnancy, postpartum period without depression, or PPD); (4) build and assess individualized ML models testing random forest, generalized linear models, support vector machine, and k-nearest neighbor algorithms; (5) validate the machine learning (ML) approach in women without PPD; (6) compare individualized model performance in women with and without PPD; (7) determine variable importance for PPD recognition; (8) generate Shapley additive explanations dependence plots to assess the relationship between digital biomarkers and PPD; and (9) compare individualized ML models versus a cohort-based model for PPD detection. EHR: electronic health record.

Figure 2

A schematic of postpartum depression (PPD) computational phenotyping.

Figure 3

Digital biomarkers vary across different periods of pregnancy among women with postpartum depression (PPD). The percentage of women in the PPD cohort exhibiting a significant difference in digital biomarker values between each pair of periods (left [represented by 0-100]) and the direction of their relationship (right). The x-axis illustrates a comparison of Tukey honest significant differences (HSD) between 2 periods of interest, representing the subtraction of digital biomarker values between the first and second periods. Tukey HSD tests were individually conducted for each woman’s data, and the percentage showing a significant relationship was calculated and presented on the heat map. The heat map on the right illustrates the overall relationship between the digital biomarker during the 2 periods of interest among the women who exhibited a significant relationship (as indicated by the percentage shown on the left heat map), with the period listed second serving as the reference. In summary, the findings indicated that digital biomarkers undergo significant alterations across different periods of pregnancy on an individual basis. Calories BMR: calories burned during the basal metabolic rate; HR: heart rate; NS: not significant; Q1: quartile 1; Q3: quartile 3.

Figure 4

Individualized machine learning models for postpartum depression (PPD) recognition outperformed those in women without PPD detecting the PPD-equivalent period. The sensitivity, specificity, precision, recall, and F1-score were calculated across individual random forest models for women in the PPD and non-PPD cohorts for the prepregnancy (A), pregnancy (B), postpartum (C), and PPD or PPD-equivalent periods (D). Individualized model performance was not significantly different regarding sensitivity, specificity, precision, recall, and F1-score for predicting the prepregnancy, pregnancy, or postpartum periods between women in the PPD and non-PPD cohorts. Individualized model performance was reduced for sensitivity, precision, recall, and F1-score, whereas specificity did not differ between the PPD and non-PPD cohorts. Data are expressed as mean and SD.

Figure 5

The direction of digital biomarkers in machine learning models for postpartum depression (PPD) classification was heterogeneous. (A) The percentage of women in the PPD cohort with a significant Pearson correlation (left) and the net relationship (right) for the top 5 overall ranked digital biomarkers for PPD classification. (B) The percentage of women in the non-PPD cohort with a significant Pearson correlation (left) and the net relationship (right) for the top 5 overall ranked digital biomarkers for PPD-equivalent classification. The proportion of women showing a significant Pearson correlation coefficient between Shapley additive explanations (SHAP) values and digital biomarkers varied in both the PPD and non-PPD cohorts. The x-axis illustrates the comparison of 2 periods, with the first period as the reference, whereas the shading indicates the percentage (0%-100%) of women showing a significant relationship (A) or the net relationship among those with a significant relationship (B). SHAP dependence plots were generated for each woman on an individual basis. For instance, the upper left tile in A presenting prepregnancy versus PPD and Calories BMR indicates the percentage of women who showed a significant correlation on SHAP dependence plots of calories burned during the basal metabolic rate (calories BMR) between the prepregnancy and PPD periods. In B, the upper left tile of the heat map for prepregnancy versus PPD and Calories BMR illustrates that most women in the PPD cohort showed a positive relationship (elevated SHAP values with increases in calories BMR, meaning that a higher level of calories BMR was more predictive of PPD than in the prepregnancy period) among women with a significant correlation, as shown in A. Among those showing a significant relationship in SHAP dependence plots during the prepregnancy versus PPD (and prepregnancy versus PPD-equivalent) periods, the correlation pattern for SHAP values and calories BMR differed—most women exhibited a positive correlation in the PPD cohort, whereas there was no uniform pattern among women in the non-PPD cohort. Among women in the pregnancy versus PPD and postpartum versus PPD (and PPD-equivalent) periods, most demonstrated a negative relationship between SHAP values and calories BMR in both the PPD and non-PPD cohorts. HR: heart rate; NS: not significant; Q1: quartile 1.