Healthy habits by chatbots : a pilot RCT study of the effects of conversational agent-based multi-domain lifestyle interventions programs on habit formation of elderly users

소민경 Graduate School, Yonsei University 2021 국내석사

전 세계적으로 치매 유병률이 증가하고 있으며, 현재 국내 65 세 이상 노인 인구 중 치매 환자 수는 약 75 만 명에 다다른다. 선행 연구에 따르면, 치매는 조기에 예방 할수록 발병률이 낮으며, 다영역 라이프스타일 중재(Multidomain lifestyle intervention)가 인지 능력 감소를 예방하는데 효과적이다. 따라서, 본 연구에서는 습관 형성 이론을 기반으로 대화형 에이전트를 활용하여 다영역 라이프스타일 프로그램을 개발하고, 이를 통해 인지 능력 강화에 대한 효과성 및 사용자의 만족도를 확인하고자 하였다. 총 40 세에서 65 세 사이의 중년 59 명을 온라인 및 오프라인으로 모집하였으며, 실험군 (n=31) 과 대조군(n=28) 에 각각 랜덤으로 할당하였다. 실험 군에게는 대화형 에이전트 ‘이다’를 통해, 대조군에게는 책의 형태로 라이프스타일 중재가 진행되었다. 모든 참가자들은 온라인으로 주관적 기억력 감퇴 설문 (SMCQ), 단축형 한국판 세계보건기구 삶의 질 척도(WHOQOL-BREF), 자가 보고 습관 척도 (SRHI)를 사전과 3 주 후, 처치가 끝난 뒤에 측정하였다. 또한, 대화형 에이전트를 사용한 실험군에게는 추가로 사후 인터뷰를 통해, 사용 경험에 대해 조사하였다. 본 연구는 지각된 인지 능력과 삶의 질에는 두 집단간의 유의미한 차이가 나타나지 않았다. 하지만, 대화형 에이전트 ‘이다’ 를 사용한 그룹이 식이 및 인지 훈련 습관 강도에 있어서 유의미하게 높게 나타났다. (t(57) = 2.70, p = .009; t(57) = 2.30, p = .025 ). 또한, 주관적 기억력은 두 그룹 모두에서 유의미하게 향상되었다 (t(30) = 2.49, p = .019; t(27)=3.33,p=.003). 또한,‘이다’ 를 사용한 그룹에서는 네 가지 습관 즉, 식이, 운동, 인지 훈련, 사회활동 모두에서 습관 강도가 사전에 비해 사후에 유의미하게 높게 나타났다 (t(30) = -6.25 to -2.44, p = .000 to .021). 결론적으로, 대화형 에이전트를 통한 라이프스타일 습관 형성은 인지 능력 감소를 줄이고, 습관 강도를 높이는 데에 효과적이다. 따라서, 향후에 더 정확한 인지 측정 도구를 사용하여 많은 수의 참가자들에게 장 기간 동안 연구를 진행할 필요가 있다. The prevalence of dementia is increasing around the world. The number of dementia patients among the nation's elderly population aged 65 or older now reaches about 750,000. According to prior research, the earlier prevention, the lower, the incidence of dementia occurs. Furthermore, many studies support the effectiveness of Multidomain lifestyle Intervention in preventing cognitive decline. In this study, we developed a multidomain lifestyle program based on habit formation theory by using a conversational agent for improving cognitive ability. We then verified its effectiveness and user satisfaction. A total of 59 middle-aged between 40-65 years participants with normal cognitive abilities were recruited online and randomly assigned to either an intervention group (n =31) or a control group (n = 28). Participants from both groups were introduced to lifestyle intervention programs. The program was provided in the form of conversational agents (Ida) for the intervention group and a textbook for the control group. All subjects completed web-based versions of the Subjective Memory Complaints Questionnaire (SMCQ), the Korean version of World Health Organization Quality of Life Assessment Instrument short form (WHOQOL- BREF), and Self-Report Habit Index (SRHI) at baseline and 3 weeks after the intervention. Besides, we conducted a follow-up interview with participants who used the conversational agent program to ask them about their experience with their use. This study showed no significant difference in the change of perceived cognitive ability and quality of life between the two groups. However, Ida group had significantly higher dietary and cognitive training habits than the control group (t(57) = 2.70, p = .009; t(57) = 2.30, p = .025 ). In addition, perceived cognitive ability was significantly improved in both groups (t(30) = 2.49, p = .019; t(27) = 3.33, p = .003). Four habit strengths such as diet, exercise, cognitive training, and social activities in the Ida group were significantly higher than the baseline (t(30) = -6.25 to -2.44, p = .000 to .021). In conclusion, the formation of lifestyle habits for preventing dementia through a conversational agent improves cognitive ability and is also effective in strengthening habit strength. In future studies, it would be better to conduct this study with precise cognitive measurement, a larger number of participants, and for a longer period.

Feasibility and Efficacy of a Community-Based, Lifestyle Weight Management Intervention Among Breast Cancer Survivors

DeScenza, Victoria R ProQuest Dissertations & Theses The Ohio State Uni 2022 해외박사(DDOD)

Objective. To establish the feasibility and efficacy of a group-mediated cognitive behavioral (GMCB) exercise and dietary (EX+D) intervention among overweight or obese breast cancer survivors (BCaS).Methods. The Healthy New Albany: Breast Cancer study was a single arm pilot trial with 21 BSC assigned to a 6-month behavioral lifestyle intervention. The exercise component of the personalized EX+D intervention integrated a combination of supervised resistance and aerobic exercise performed twice per week. The dietary component involved counseling and education to modify dietary intake and composition.Results. Feasibly parameters yielded generally favorable recruitment, retention, and adherence rates, together with the absence of intervention-related adverse events. Effect size calculations were used to examine the potential effects that the intervention had on changes in body composition and select social cognitive outcomes at 3-month and 6-month follow-up. Meaningful changes were observed in total body mass (d=0.22), fat mass (d= 0.40), percent fat mass (d= 0.51), percent lean mass (d= 0.63), and upper body muscular strength (d= 0.95). Additional positive effects were observed in perceived competence of diet (3M: d=-0.46 and 6M d=-1.19) and improvement in perceived competence of exercise (3-Month: d=-0.16 and 6-Month: d=-.34). Furthermore, self-monitoring (d= -1.16), goal setting (d=-0.62), and reinforcement (d=-0.27) saw positive effects at 6-months.Conclusions: The GMCB LWS intervention is safe and well tolerated by overweight or obese breast cancer survivors and yielded favorable improvements in relevant objective and patient reported social cognitive outcomes among BCaS.

Development and feasibility evaluation of a chatbot-based lifestyle intervention for pre-menopausal women

김연지 Graduate School, Yonsei University 2023 국내박사

본 연구는 대사적으로 건강하지 않은 정상체중(Metabolically Unhealthy Normal Weight, MUNW)인 폐경 전 여성을 대상으로 챗봇 기반 생활습관 프로그램 개발과 실현 가능성 평가를 목적으로 한다. 챗봇 기반 생활습관 프로그램은 ADDIE 교육 모형의 다섯 단계, 분석, 설계, 개발, 구현, 평가의 과정을 통해 개발되었다. 개발된 프로그램은 델파이 기법을 통하여 전문가 타당도 검증을 받았다. 단일 그룹 사전 사후 비교 설계 연구를 수행하였다. 2022년 10월부터 2023년 3월에 일개 상급종합병원의 건강검진센터를 방문한 여성 수검자 7,455명을 스크리닝하여 선정기준에 부합하는 145명 중 연구참여자를 모집했고, 이 중 35명이 동의하여 6주 중재를 시행하였다. 프로그램에 대한 실현 가능성은 대상자 모집율, 대상자의 프로그램 완료율, 실제 프로그램의 사용정도, 만족도 및 인지된 유용성으로 평가했다. 중재 전과 후에 자가설문을 이용하여 대사증후군 관련 지식과 인식도, 건강태도, 건강관리 자기효능감을 측정했고, 대사건강 관련 자료는 신세계측, 혈압과 혈액검사를 통해 수집했다 (체질량지수, 체지방률, 복부지방률, 허리둘레, 혈압, 고밀도지단백 콜레스테롤, 중성지방). 연구에 참여한 35명 중 1명을 제외한 34명이 MUNW에 대한 궁금증과 관리법이 궁금하여 연구에 참여하였고 응답했다. 35명 중 68.6%가 챗봇에 대해서 만족하였고, 85.7%가 건강생활습관 프로그램에 대해 만족한다고 보고했다. 특히, MUNW에 대한 정의와 합병증에 관한 콘텐츠가 유용하다고 응답하였다. 중재 전 ∙ 후 혈액검사 결과를 비교한 자료에 의하면 MUNW 관련 기준 중 공복 혈당 관련 요인이 유의하게 감소하였고(p = <0.05), 건강관리 자기효능감은 70.69±12.08에서 89.88±14.54점으로 통계적으로 유의하게 증가하였다(p = <0.001). 본 연구에서 개발한 챗봇 기반 생활습관 프로그램은 폐경 전 MUNW 여성의 대사건강상태에 대한 인식을 높이고 건강한 건강관리 수행을 돕고자 정보 – 동기 – 행동기술 모델을 적용하고, 이를 상용화한 소셜 미디어 플랫폼의 챗봇 기능을 활용하여 설계했다. 본 연구의 대상이 일 대학병원의 건강검진 센터의 검진 대상자로 제한되었고, 표본수가 많지 않기 때문에 주의 깊은 해석이 필요하다. 문헌고찰과 다학제간 전문가들의 의견을 바탕으로 한 교육자료를 개발하고, 이를 대상자들에게 적용하여 이들의 대사건강에 대한 인식, 생활습관 개선에 대한 가능성을 탐색함으로써 프로그램의 실현 가능성을 파악하였다. 향후 대상자 수 확대, 챗봇 구조 및 콘텐츠를 확충하여 반복 연구를 통해 광범위한 지역사회 수준에서 폐경 전 여성들의 대사건강에 대한 인식을 높이고 생활습관 개선에 동기를 부여하는 중재방법으로 활용될 것으로 기대한다. Background: For premenopausal women, prioritizing healthy lifestyle is more important to promote metabolic health status than solely focusing on maintaining a healthy body mass index (BMI). Objectives: The purpose of this study was to develop a healthy lifestyle program for premenopausal women who are in the state of metabolically unhealthy normal weight (MUNW) and to assess the feasibility of the program. Design and Setting: This study involved a methodological study for program development and a prospective, a single group pre- and posttest intervention study for program feasibility evaluation. This study was conducted at a health checkup center affiliated with a tertiary academic medical center. Participants: Study participants were 35premenopausal women with MUNW, whose BMI is less than 25kg/m2, but who presented with a combination of two or more of the following metabolic risks: (1) systolic blood pressure (SBP) > 130 mmHg or diastolic blood pressure (DBP) > 85 mmHg, (2) fasting blood sugar (FBS) ≥ 100 mg/dL, (3) HDL-C < 50 mg/dL, (4) Triglyceride ≥ 150 mg/dL. Method: A 6-week chatbot based healthy lifestyle program was developed according to the five stages of the Analysis, Design, Development, Implementation, and Evaluation (ADDIE) instructional model. Addressing key components of the Information-Motivation-Behavioral skills (IMB) model, the intervention contents consisted of basic information on MUNW and guides to motivate practicing healthy lifestyle (i.e., physical activity, diet). To assess feasibility, we monitored recruitment and retention rates, and participant-rate program satisfaction and usefulness. Health behavior information, health attitude and health management self-efficacy were assessed before and after the program. BMI and metabolic health status related measures (i.e., SBP, DBP, FBS, HDL-C, and TG) were obtained at before enrollment and after completion of the 6-week program. Results: From October 2022 to March 2023, after screening a total of 7,455 women, 145 eligible women were approached for study participation and 35 women agreed to consent and completed the baseline survey. All participants completed the 6-week intervention program and completed post-test measures. Regarding the participant overall satisfaction to the program, a majority of participants reported either very satisfied (n = 12. 34.3%) or satisfied (n = 18, 51.4%) with the program. Most participants positively rated their perceived usefulness of each topic, especially all topic on the definition of MUNW and complications of MUNW. While there were no significant changes in the measures of metabolic health status related variables from pre- to post-intervention, there was a significant change in the proportion of participants with glucose-related risk factors from pre (n=13, 37.1%) to post-intervention (n=6, 17.4%, McNemar test, p < 0.05). Health self-efficacy was significantly improved from pre- to post-intervention (p < 0.001). The scores of health attitude and health promotion reported trends of improvement but the changes were not statistically significant. Conclusions: Findings of this study demonstrates this chatbot-based health behavior program is feasible and acceptable in premenopausal women with MUNW. Future studies are warranted to implement and evaluate this program in a larger and diverse sample of community-dwelling premenopausal women.

Effects of a brief intervention for cognitive health among community elderly : a cluster randomized controlled trial

이강수 Graduate School, Yonsei University 2011 국내박사

Background: There is now widespread public interest in developing strategies to maintain or enhance cognitive health in the elderly. Cognitive decline in the elderly involves multiple pathophysiologic and psychosocial processes. Thus, research that focuses on preserving cognition may well identify a different set or combination of risk factors and thus different prevention strategies for healthy elderly subjects.Methods: This study examines 12-month outcomes from a clustering randomized controlled trial of a 5 group intervention vs a control condition among elderly aged over 60 years old. Hypotheses were that the interventions would be more effective than the control in reducing cognitive decline at 12 months’ follow-up.Results: This report describes a cluster randomized controlled trial of a brief intervention in a group of community dwelling elderly. This is the first study to examine the effectiveness of an intervention for cognitive health in a community setting. We found that the trial resulted in a change of mental and social activity in the score for the intervention group compared with the control group. However, the intervention did not result in a clinically meaningful improvement in MMSE scores.Conclusion: Considering that lifestyle behaviors co-occur, multiple health behavior approach appears more practical. In the brief intervention, encouragement of mental and social activity would be helpful in cognitive health of the community dwelling elderly.

저소득 중년여성을 위한 라이프스타일 중재가 건강증진행위, D 유형 성격, 인지기능과 신체조성에 미치는 효과

김은진 전북대학교 일반대학원 2021 국내석사

Reducing Blood Pressure Using an Educational Intervention Related to Lifestyle Modification Among African American Adults

Sule, Ololade ProQuest Dissertations & Theses Brandman Universit 2020 해외박사(DDOD)

Hypertension is a chronic disease that represents a significant public health challenge both in the United States and across the world. While the incidence of hypertension cuts across all ethnic groups, the disease affects the African American population disproportionately and has become one of the leading causes of morbidity and mortality within this racial group. The high mortality rate associated with the complications of uncontrolled hypertension is an ongoing problem and continues to be a concern despite efforts to tackle the issue. Effective management of uncontrolled hypertension can reduce morbidity and mortality. The aim of this clinical scholarly project was to establish a model that could reduce blood pressure among African American adults with uncontrolled hypertension. The intervention was designed to promote and encourage lifestyle modifications. For the project, a survey was designed to obtain pre- and post-treatment data from participants about their knowledge of hypertension, its complications, and its treatment options. A convenience sample of 20 African American adults with uncontrolled hypotension was taken. The results of the project indicated that educational lifestyle intervention was effective in significantly decreasing participants’ systolic (p < .001) and diastolic (p = .022) blood pressure.

페루의 도시 저소득 지역에서의 고혈압관리를 위한 지역사회 기반 중재효과 평가

조혜연 연세대학교 대학원 2020 국내박사

The prevalence of hypertension and non-communicable diseases(NCDs) in developing countries is high, but the rates of awareness, treatment and control of hypertension are lower compared to the rates of high-income countries. In Peru, although the burden of high blood pressure and NCDs is great, studies on interventions for the blood pressure(BP) management are poor. This study aims to assess the effectiveness of a community-based intervention for controlling hypertension in low-income urban areas in Peru implemented by Korea International Cooperation Agency(KOICA), and to provide an evidence for further implementation of the intervention model in similar areas. For this study, we used a non-equivalent control group pretest-posttest design. Among the study subjects, 340 people were selected as experimental group who completed a 12-week community-based intervention for controlling hypertension from April to December 2017, while 317 people were selected as control group who only participated in pre- and post- need assessment during the same period. Same standards of selection were applied on both experimental and control group, and we selected adults aged 18 years or older and in a pre-hypertension or hypertension state. The intervention program consisted of individualized consultations, group health education sessions, regular self-help group meetings, and text messages providing health education was applied to the experimental group. A difference-in-differences regression analysis was conducted to evaluate the effectiveness of the intervention on the level of hypertension-related knowledge, level of practicing health behavior, body mass index(BMI), waist circumference, systolic blood pressure, diastolic blood pressure and the rate of blood pressure control in the experimental group. Logistic regression analysis were also performed to analyze factors affecting the blood pressure control in the experimental group. As results, the level of hypertension-related knowledge improved significantly in the experimental group rather than in the control group(p<.001). Level of practicing health behavior related to eating habits, stress management and medication compliance has significantly increased in the experimental group over the control group (p<.05). However, the change of experimental group in smoking, drinking and practice of physical activity were not statistically significant compared to the control group. The rate of blood pressure control was improved(p<.01) and systolic blood pressure(p<.001) and waist circumference(p<.05) was decreased in experimental group over the control group. However, the changes in BMI and diastolic blood pressure in experimental group were not statistically significant compared to the control group. Before the intervention, factors affecting blood pressure control of experimental group were sex, age and the use of anti-hypertensive medication. After the intervention, sex, age, use of anti-hypertensive medication and eating habits significantly affected on the blood pressure control of experimental group. In conclusion, we verified that the community-based intervention for controlling hypertension implemented in the low-income urban area in Peru was effective in improving hypertension-related knowledge level, eating habits, stress management, adherence to anti-hypertensive medication, decreasing waist circumference and systolic blood pressure and controlling the blood pressure of the participants. Therefore, it is necessary to expand the implementation of intervention model to other similar areas in Peru. For this, it is suggested to establish relevant regulations at the regional health department level, supplement health personnel and build their capacity on disease prevention and health promotion and create a environment to promote physical activity of the community people. 개발도상국에서 고혈압과 합병증인 만성질환의 발생률은 높은 수준이지만 고혈압 인지율, 치료율 및 조절률은 고소득국가들보다 낮다. 남미 지역 국가인 페루 역시 고혈압으로 인한 질병 부담이 증가하고 있으나 고혈압관리 중재에 관해 수행된 지역 내 연구는 부족한 실정이다. 이에 본 연구에서는 페루의 도시 저소득 지역에서 수행된 지역사회 기반 KOICA 고혈압관리 중재의 효과를 평가하고, 향후 중재 모형을 확대 적용하기 위한 근거를 제시하고자 하였다. 본 연구에서는 비동등성 대조군 전후 비교 유사실험 설계를 사용하였다. 연구대상자는 고혈압 또는 고혈압 전단계 상태인 18세 이상의 성인으로, 실험군은 2017년 4월부터 12월까지 9개월 동안 3개의 실험군 보건소에서 고혈압관리 중재를 이수한 340명, 대조군은 같은 기간에 2개의 대조군 보건소에서 사전조사와 사후조사에 참여한 317명이었다. 중재 프로그램은 12주 과정으로 개별 상담, 집단 보건교육, 자조모임, 문자메시지 전송을 통한 혈압관리 행동지침 전달 등으로 구성되었다. 중재 참여자의 건강지식 수준, 건강행동 수행정도, 체질량지수, 허리둘레, 수축기혈압, 이완기혈압 및 혈압 조절 여부에 대한 중재의 효과성을 확인하기 위해 이중차이분석을 실시하였고, 중재 참여자의 혈압 조절에 미치는 영향요인을 파악하기 위해 로지스틱 회귀분석을 실시하였다. 연구결과, 고혈압 관련 지식, 식생활 관련 지식, 신체활동 관련 지식은 중재 후 대조군보다 실험군에서 통계적으로 유의하게 더 큰 폭으로 증가하였다(p<.001). 둘째, 식습관, 스트레스 관리, 복약순응 관련 건강행동 수행 정도는 중재 후 대조군보다 실험군에서 통계적으로 유의하게 더 큰 폭으로 증가하였다(p<.05). 신체활동 실천, 금연, 절주 관련 건강행동 수행 정도는 중재 전보다 후에 지표가 개선되었으나 통계적으로 유의하지는 않았다. 셋째, 혈압 조절 여부(p<.01)와 수축기혈압(p<.001), 허리둘레(p<.05)는 중재 후 대조군보다 실험군에서 통계적으로 유의하게 더 큰 폭으로 개선되었다. 그러나 체질량지수와 이완기혈압에서는 중재 후 유의미한 변화가 나타나지 않았다. 넷째, 혈압 조절에 미치는 영향요인을 분석한 결과 중재 전에는 성과 연령, 혈압약 복용 여부가 혈압 조절 여부에 통계적으로 유의한 영향을 미치는 것으로 나타났고, 중재 후에는 성, 연령, 혈압약 복용 여부와 식습관 점수가 유의미한 영향을 미치는 것으로 나타났다. 결론적으로 페루의 도시 저소득 지역에서 수행한 지역사회 기반 고혈압관리 중재는 참여자의 고혈압 관련 지식 수준과 식습관, 스트레스 관리, 복약순응에 관한 건강행동 수행정도, 허리둘레 감소, 수축기혈압 감소 및 혈압 조절률 향상에 효과가 있음을 확인하였다. 따라서 향후 페루 내 유사 지역에 고혈압관리 중재 모형을 확대 시행할 필요가 있으며, 이를 위해서는 건강증진 관련 보건의료인력 확충과 역량강화, 신체활동 증진을 위한 지역사회 환경 조성과 지자체 또는 지역보건국 차원의 사업 시행 관련 규정 및 계획 수립이 필요할 것이다.

(A) prediction model for prevention and management of metabolic syndrome based on machine learning

이정훈 Graduate School, Yonsei University 2023 국내박사

Digital health-based lifestyle interventions (e.g., mobile applications, short message services, wearable devices, social media, and interactive websites) are widely used to manage metabolic syndrome (MetS). This study aimed to confirm the usefulness of digital health-based lifestyle interventions using healthcare devices and propose a novel prediction model of prevention and management for MetS. Participants with one or more MetS risk factors were recruited from December 2019 to September 2020, and finally, 106 participants were analyzed. Participants were provided with five healthcare devices and applications. Characteristics were compared at baseline and follow-up, and lifelog data that were collected during the clinical trial were analyzed. With these results, the frequency of use of healthcare devices for continuous self-care was quantified, and a novel prediction model for the prevention and management of MetS was developed. The model predicts persistence in continuous engagement as well as abbreviated risk factors for self-care effects. Representative machine-learning classifiers were used and compared. In both models, the random forest classifier showed the best performance, and feature selection was optimized through random forest-recursive feature elimination. As a result, the prediction model for persistence showed recall of 83.0%, precision of 92.4%, an F1-score of 0.874, a Matthews correlation coefficient (MCC) of 0.844, and accuracy of 94.9%. The prediction model for abbreviated risk factors showed a recall of 79.8%, a precision of 87.2%, an F1-score of 0.834, and an MCC of 0.797 for increased abbreviated risk factors, and a recall of 75.1%, a precision of 85.5%, an F1-score of 0.800, and an MCC of 0.747 for decreased abbreviated risk factors. The prediction model proposed showed high performance. Based on self-care with digital health-based lifestyle interventions, prediction models could be helpful for the prevention and management of MetS.

An Examination of Intervention Components that Influence Engagement in Latino Community-Based Diabetes Prevention Program

Doucet, Shanna Arizona State University ProQuest Dissertations & 2023 해외박사(DDOD)

This study examined the responses of Latino community residents diagnosed with pre-diabetes using a complex convergent mixed methods design nested within a three-factor factorial design (age, language, and gender) to understand the intervention contents and activities that are most important to study participants for helping them to prevent the development and onset of type 2 diabetes (T2D). Chapter 1 introduces the topic and provides background literature. Chapter 2 systematically synthesized findings from diabetes prevention studies to identify efficacious intervention components and social contextual resources that have been used to support diabetes prevention in Latinos who are prediabetic or at risk of developing diabetes. Chapter 3 analyzed the qualitative portion of one section within the Latino Lifestyles study protocol. Based on the Integrative Mixed Methods (IMM) methodology, we conducted a thematic analysis to identify emergent themes for each of the three focus questions, as administered in individual interviews with 28 Latino and Latina participants. Chapter 4 utilized the findings from Chapter 3 to examine the levels and associations of three potential moderator variables: (a) health motivation, (b) diabetes awareness, and (c) diabetes concern among Latinos diagnosed with prediabetes. A 2x2 cross-tabulation analysis tested group differences in the mention of response phrases as examined by two levels (high versus low) of the potential moderator variables. The study results offered meaningful intervention components such as healthy cooking classes and assistance with lifestyle changes related to diet and exercise for themselves and their families, as expressed by the participants. These findings provide informative intervention components for developing a Latino community-based diabetes prevention program designed to enhance engagement and lifestyle changes for individuals diagnosed with prediabetes.

Digital Tools to Enable Large-Scale Access to Biomechanical Assessment

Boswell, Melissa Ann ProQuest Dissertations & Theses Stanford Universit 2022 해외박사(DDOD)

How we move and how much we move profoundly affect our health and wellbeing. In the field of biomechanics, we analyze movement to help identify, monitor, and treat diseases and disorders that impact movement. However, biomechanical intervention and monitoring have historically been restricted to a laboratory setting. This dissertation proposes an interdisciplinary approach to improving accessibility to biomechanical intervention and monitoring through bridging advancements in computer science, biomechanics, and psychology. Through three studies, I describe the development of new machine learning methods, mobile biomechanical tools, and digital psychological intervention to improve access to biomechanical assessments.In the first study, we developed a machine learning model to predict a key biomechanical measure of osteoarthritis progression. Altering the foot progression angle is a gait modification that aims to reduce the knee adduction moment, a surrogate measure of internal knee loading, to improve pain and slow disease progression in individuals with knee osteoarthritis. Foot progression angle modifications must be personalized to an individual, which traditionally requires an expensive gait laboratory. In this study, I demonstrate the feasibility of identifying personalized biomechanical interventions with a smartphone camera. We developed a machine learning model to predict the peak knee adduction moment using anatomical landmark positions identifiable with two-dimensional video. The model was 92% accurate in predicting changes in the peak knee adduction moment with foot progression angle modification, suggesting that it is feasible to use this tool with smartphone camera videos to prescribe personalized gait modifications in clinical settings.If made accessible for use with only a smartphone, pose estimation algorithms that predict joint locations from two-dimensional video have the potential to be used for at-home biomechanical self-assessment. In the second study, we tested the clinical relevance of pose estimation from at-home smartphone videos with the five-repetition sitto-stand test (5STS), one of the most widely used tests for measuring an individual's physical functioning. We created a web app that gave participants instructions to upload and record an at-home video of the 5STS, which was automatically processed to extract 5STS timing and kinematics. After remotely collecting data from over 350 participants, we evaluated the relationships between automatically-extracted 5STS timing and kinematics and self-reported demographic and health data. We found that longer times to complete the 5STS test were associated with older age, higher BMI, and lower physical health score. A larger maximum trunk flexion angle was associated with the presence of osteoarthritis. Trunk flexion angle remained a predictor of osteoarthritis presence (p=0.037) even when controlling for age, sex, and BMI. Our large dataset enabled exploratory analyses that found associations between trunk flexion angle and ethnicity and greater trunk acceleration and higher mental health score. The consistency of our results with lab-based studies supports the use of our open-source web app by researchers and clinicians to leverage biomechanics for at-home monitoring of physical functioning.Increasing access to movement-based interventions does not guarantee that individuals will use them or increase their activity levels because of them. Exercise is important for improving knee pain and functioning in individuals with osteoarthritis. However, exercise participation and adherence to physical activity interventions are low in the osteoarthritis population.