Automatically detecting activities of daily living from in-home sensors as indicators of routine behaviour in an older population

Objective

The NEX project has developed an integrated Internet of Things (IoT) system coupled with data analytics to offer unobtrusive health and wellness monitoring supporting older adults living independently at home. Monitoring involves visualising a set of automatically detected activities of daily living (ADLs) for each participant. ADL detection allows the incorporation of additional participants whose ADLs are detected without system re-training.

Methods

Following a user needs and requirements study involving 426 participants, a pilot trial and a friendly trial of the deployment, an action research cycle (ARC) trial was completed. This involved 23 participants over a 10-week period each with ∼ 20 IoT sensors in their homes. During the ARC trial, participants took part in two data-informed briefings which presented visualisations of their own in-home activities. The briefings also gathered training data on the accuracy of detected activities. Association rule mining was used on the combination of data from sensors and participant feedback to improve the automatic ADL detection.

Results

Association rule mining was used to detect a range of ADLs for each participant independently of others and then used to detect ADLs across participants using a single set of rules for each ADL. This allows additional participants to be added without the necessity of them providing training data.

Conclusions

Additional participants can be added to the NEX system without the necessity to re-train the system for automatic detection of their ADLs.

Development of an Internet of Things Technology Platform (the NEX System) to Support Older Adults to Live Independently: Protocol for a Development and Usability Study

Background

In a rapidly aging population, new and efficient ways of providing health and social support to older adults are required that not only preserve independence but also maintain quality of life and safety.

Objective

The NEX project aims to develop an integrated Internet of Things system coupled with artificial intelligence to offer unobtrusive health and wellness monitoring to support older adults living independently in their home environment. The primary objective of this study is to develop and evaluate the technical performance and user acceptability of the NEX system. The secondary objective is to apply machine learning algorithms to the data collected via the NEX system to identify and eventually predict changes in the routines of older adults in their own home environment.

Methods

The NEX project commenced in December 2019 and is expected to be completed by August 2022. Mixed methods research (web-based surveys and focus groups) was conducted with 426 participants, including older adults (aged ≥60 years), family caregivers, health care professionals, and home care workers, to inform the development of the NEX system (phase 1). The primary outcome will be evaluated in 2 successive trials (the Friendly trial [phase 2] and the Action Research Cycle trial [phase 3]). The secondary objective will be explored in the Action Research Cycle trial (phase 3). For the Friendly trial, 7 older adult participants aged ≥60 years and living alone in their own homes for a 10-week period were enrolled. A total of 30 older adult participants aged ≥60 years and living alone in their own homes will be recruited for a 10-week data collection period (phase 3).

Results

Phase 1 of the project (n=426) was completed in December 2020, and phase 2 (n=7 participants for a 10-week pilot study) was completed in September 2021. The expected completion date for the third project phase (30 participants for the 10-week usability study) is June 2022.

Conclusions

The NEX project has considered the specific everyday needs of older adults and other stakeholders, which have contributed to the design of the integrated system. The innovation of the NEX system lies in the use of Internet of Things technologies and artificial intelligence to identify and predict changes in the routines of older adults. The findings of this project will contribute to the eHealth research agenda, focusing on the improvement of health care provision and patient support in home and community environments.

International Registered Report Identifier (IRRID)

DERR1-10.2196/35277

Respondent Characteristics and Dietary Intake Data Collected Using Web-Based and Traditional Nutrition Surveillance Approaches: Comparison and Usability Study

Background

There are many constraints to conducting national food consumption surveys for national nutrition surveillance, including cost, time, and participant burden. Validated web-based dietary assessment technologies offer a potential solution to many of these constraints.

Objective

This study aims to investigate the feasibility of using a previously validated, web-based, 24-hour recall dietary assessment tool (Foodbook24) for nutrition surveillance by comparing the demographic characteristics and the quality of dietary intake data collected from a web-based cohort of participants in Ireland to those collected from the most recent Irish National Adult Nutrition Survey (NANS).

Methods

Irish adult participants (aged ≥18 years) were recruited to use Foodbook24 (a web-based tool) between March and October 2016. Demographic and dietary intake (assessed by means of 2 nonconsecutive, self-administered, 24-hour recalls) data were collected using Foodbook24. Following the completion of the study, the dietary intake data collected from the web-based study were statistically weighted to represent the age-gender distribution of intakes reported in the NANS (2008-2010) to facilitate the controlled comparison of intake data. The demographic characteristics of the survey respondents were investigated using descriptive statistics. The controlled comparison of weighted mean daily nutrient intake data collected from the Foodbook24 web-based study (329 plausible reporters of a total of 545 reporters) and the mean daily nutrient intake data collected from the NANS (1051 plausible reporters from 1500 reporters) was completed using the Wilcoxon–Mann-Whitney U test in Creme Nutrition software.

Results

Differences between the demographic characteristics of the survey participants across the 2 surveys were observed. Notable differences included a lower proportion of adults aged ≥65 years and a higher proportion of females who participated in the web-based Foodbook24 study relative to the NANS study (P<.001). Similar ranges of mean daily intake for the majority of nutrients and food groups were observed (eg, energy [kilocalorie per day] and carbohydrate [gram per day]), although significant differences for some nutrients (eg, riboflavin [mg/10 MJ], P<.001 and vitamin B12 [µg/10 MJ], P<.001) and food groups were identified. A high proportion of participants (200/425, 47.1%) reported a willingness to continue using Foodbook24 for an additional 6 months.

Conclusions

These findings suggest that by using targeted recruitment strategies in the future to ensure the recruitment of a more representative sample, there is potential for web-based methodologies such as Foodbook24 to be used for nutrition surveillance efforts in Ireland.

Physical, psychological and nutritional outcomes in a cohort of Irish patients with metastatic peritoneal malignancy scheduled for cytoreductive surgery (CRS) and heated intrapertioneal chemotherapy (HIPEC): An exploratory pilot study

Background

Treatment for peritoneal malignancy (PM) can include cytoreductive surgery (CRS) and heated intrapertioneal chemotherapy (HIPEC) and is associated with morbidity and mortality. Physical, psychological and nutritional outcomes are important pre-operatively. The aim of this pilot study was to investigate these outcomes in patients with PM before and after CRS-HIPEC.

Methods

Between June 2018 and November 2019, participants were recruited to a single-centre study. Primary outcome was cardiopulmonary exercise testing (CPET) variables oxygen uptake (VO₂) at anaerobic threshold (AT) and at peak. Secondary outcome measures were upper and lower body strength, health related quality of life (HRQoL) and the surgical fear questionnaire. Exploratory outcomes included body mass index, nutrient intake and post-operative outcome. All participants were asked to undertake assessments pre CRS-HIPEC and 12 weeks following the procedure.

Results

Thirty-nine patients were screened, 38 were eligible and 16 were recruited. Ten female and 6 male, median (IQR) age 53 (42–63) years. Of the 16 patients recruited, 14 proceeded with CRS-HIPEC and 10 competed the follow up assessment at week 12. Pre-operative VO₂ at AT and peak was 16.8 (13.7–18) ml.kg^-1.min^-1 and 22.2 (19.3–25.3) ml.kg^-1.min^-1, upper body strength was 25.9 (20.3–41.5) kg, lower body strength was 14 (10.4–20.3) sec, HRQoL (overall health status) was 72.5 (46.3–80) % whilst overall surgical fear was 39 (30.5–51). The VO₂ at AT decreased significantly (p = 0.05) and HRQoL improved (p = 0.04) between pre and post- CRS-HIPEC. There were no significant differences for any of the other outcome measures.

Conclusion

This pilot study showed a significant decrease in VO₂ at AT and an improvement in overall HRQoL at the 12 week follow up. The findings will inform a larger study design to investigate a prehabilitation and rehabilitation cancer survivorship programme.

Dairy Consumption and Metabolic Health

Milk and dairy foods are naturally rich sources of a wide range of nutrients, and when consumed according to recommended intakes, contribute essential nutrients across all stages of the life cycle. Seminal studies recommendations with respect to intake of saturated fat have been consistent and clear: limit total fat intake to 30% or less of total dietary energy, with a specific recommendation for intake of saturated fat to less than 10% of total dietary energy. However, recent work has re-opened the debate on intake of saturated fat in particular, with suggestions that recommended intakes be considered not at a total fat intake within the diet, but at a food-specific level. A large body of evidence exists examining the impact of dairy consumption on markers of metabolic health, both at a total-dairy-intake level and also at a food-item level, with mixed findings to date. However the evidence suggests that the impact of saturated fat intake on health differs both across food groups and even between foods within the same food group such as dairy. The range of nutrients and bioactive components in milk and dairy foods are found in different levels and are housed within very different food structures. The interaction of the overall food structure and the nutrients describes the concept of the ‘food matrix effect’ which has been well-documented for dairy foods. Studies show that nutrients from different dairy food sources can have different effects on health and for this reason, they should be considered individually rather than grouped as a single food category in epidemiological research. This narrative review examines the current evidence, mainly from randomised controlled trials and meta-analyses, with respect to dairy, milk, yoghurt and cheese on aspects of metabolic health, and summarises some of the potential mechanisms for these findings.

Generic Meal Patterns Identified by Latent Class Analysis: Insights from NANS (National Adult Nutrition Survey)

Nutritional data reduction methods are widely applied in nutrition epidemiology in order to classify individuals into meaningful groups with similar dietary patterns. To date, none of the existing studies have applied latent class analysis to examine dietary patterns which include meal types consumed throughout a day. We investigated main meal patterns followed on weekend and weekdays, and evaluated their associations with cardio-metabolic biomarkers. The analyses were performed within the NANS (National Adult Nutrition Survey) a cross-sectional national food consumption survey of 1500 nationally representative Irish adults. A total number of seven dietary patterns were identified using latent class analysis. The typical meal pattern followed by the majority of the population was characterized by consumption of cereal or toast for breakfast, skipping or consuming a sandwich for light meal, and meat or fish with potatoes, pasta or vegetables for the main meal. Eating patterns differed on weekends, and those participants who consumed meat and eggs for breakfast instead of breakfast cereal and skipped light meal were more likely to have an unhealthier dietary pattern, a higher diastolic blood pressure, and increased serum ferritin. The application of data reduction techniques to simplify the multifaceted nature of dietary data is a useful approach to derive patterns, which might shed further light on the typical dietary patterns followed by populations.

A Comparison of Food Portion Size Estimation by Older Adults, Young Adults and Nutritionists

OBJECTIVES

To investigate the ability of older adults, younger adults and nutritionists to assess portion size using traditional methods versus a computer-based method. This was to inform the development of a novel dietary assessment method for older adults "The NANA system".

DESIGN

Older and younger adults assessed the portion size of self-served portions of foods from a buffet style set up using traditional and computerised portion size assessment aids. Nutritionists assessed the portion size of foods from digital photographs using computerised portion size aids. These estimates were compared to known weights of foods using univariate analyses of covariance (ANCOVA).

SETTING

The University of Sheffield, United Kingdom.

SUBJECTS

Forty older adults (aged 65 years and over), 41 younger adults (aged between 18 and 40 years) and 25 nutritionists.

RESULTS

There was little difference in the abilities of older and younger adults to assess portion size using both assessment aids with the exception of small pieces morphology. Even though the methods were not directly comparable among the test groups, there was less variability in portion size estimates made by the nutritionists.

CONCLUSION

Older adults and younger adults are similar in their ability to assess food portion size and demonstrate wide variability of estimation compared to the ability of nutritionists to estimate portion size from photographs. The results suggest that the use of photographs of meals consumed for portion size assessment by a nutritionist may improve the accuracy of dietary assessment. Improved portion size assessment aids are required for all age groups.

The Development, Validation, and User Evaluation of Foodbook24: A Web-Based Dietary Assessment Tool Developed for the Irish Adult Population

BACKGROUND

The application of technology in the area of dietary assessment has resulted in the development of an array of tools, which are often specifically designed for a particular country or region.

OBJECTIVE

The aim of this study was to describe the development, validation, and user evaluation of a Web-based dietary assessment tool "Foodbook24."

METHODS

Foodbook24 is a Web-based, dietary assessment tool consisting of a 24-hour dietary recall (24HDR) and food frequency questionnaire (FFQ) alongside supplementary questionnaires. Validity of the 24HDR component was assessed by 40 participants, who completed 3 nonconsecutive, self-administered 24HDR using Foodbook24 and a 4-day semi-weighed food diary at separate time points. Participants also provided fasted blood samples and 24-hour urine collections for the identification of biomarkers of nutrient and food group intake during each recording period. Statistical analyses on the nutrient and food group intake data derived from each method were performed in SPSS version 20.0 (SPSS Inc). Mean nutrient intakes (and standard deviations) recorded using each method of dietary assessment were calculated. Spearman and Pearson correlations, Wilcoxon Signed Rank and Paired t test were used to investigate the agreement and differences between the nutritional output from Foodbook24 (test method) and the 4-day semi-weighed food diary (reference method). Urinary and plasma biomarkers of nutrient intake were used as an objective validation of Foodbook24. To investigate the user acceptability of Foodbook24, participants from different studies involved with Foodbook24 were asked to complete an evaluation questionnaire.

RESULTS

For nutrient intake, correlations between the dietary assessment methods were acceptable to very good in strength and statistically significant (range r=.32 to .75). There were some significant differences between reported mean intakes of micronutrients recorded by both methods; however, with the exception of protein (P=.03), there were no significant differences in the reporting of energy or macronutrient intake. Of the 19 food groups investigated in this analysis, there were significant differences between 6 food groups reported by both methods. Spearman correlations for biomarkers of nutrient and food group intake and reported intake were similar for both methods. A total of 118 participants evaluated the acceptability of Foodbook24. The tool was well-received and the majority, 67.8% (80/118), opted for Foodbook24 as the preferred method for future dietary intake assessment when compared against a traditional interviewer led recall and semi-weighed food diary.

CONCLUSIONS

The results of this study demonstrate the validity and user acceptability of Foodbook24. The results also highlight the potential of Foodbook24, a Web-based dietary assessment method, and present a viable alternative to nutritional surveillance in Ireland.

Comparison of a Web-Based 24-h Dietary Recall Tool (Foodbook24) to an Interviewer-Led 24-h Dietary Recall

Web-based tools have the potential to reduce the cost of dietary assessment; however, it is necessary to establish their performance compared to traditional dietary assessment methods. This study aims to compare nutrient and food intakes derived from Foodbook24 to those obtained from an interview-led 24-h dietary recall (24HDR). Seventy-nine adult participants completed one self-administered 24HDR using Foodbook24 and one interviewer-led 24HDR on the same day. Following a 10 days wash-out period the same process was completed again in opposite order to the previous study visit. Statistical analysis including Spearman’s rank order correlation, Mann-Whitney U tests, cross-classification analysis, and “Match”, “Omission”, and “Intrusion” rates were used to investigate the relationship between both methods. Strong, positive correlations of nutrient intake estimated using both methods was observed (r_s = 0.6–1.0; p < 0.001). The percentage of participants classified into the same tertile of nutrient intake distribution using both methods ranged from 58% (energy) to 82% (vitamin D). The overall match rate for food intake between both methods was 85%, while rates for omissions and intrusions were 11.5% and 3.5%, respectively. These results, alongside the reduced cost and participant burden associated with Foodbook24, highlight the tool’s potential as a viable alternative to the interviewer-led 24HDR.