Systematic Appraisal

Systematic Appraisal
Makroum, M. A., Adda, M., Bouzouane, A., & Ibrahim, H. (2022). Machine Learning and Smart Devices for Diabetes Management: Systematic Review. Sensors, 22(5), 1843. https://doi.org/10.3390/s22051843
APPENDIX F
Appraisal Guide
Findings of a Quantitative Study
Citation:
___________________________________________________________________________
___________________________________________________________________________
___________________________________________________________________________
Synopsis
What was the purpose of the study (research questions, purposes, and hypotheses)?
How was the sample obtained?
What inclusion or exclusion criteria were used?
Who from the sample actually participated or contributed data (demographic or clinical profile and dropout rate)?
What methods were used to collect data (e.g., sequence, timing, types of data, and measures)?
Was an intervention tested?  Yes   No
1. How was the sample size determined?
2. Were patients randomly assigned to treatment groups?
What are the main findings?
Credibility
Is the study published in a source
that required peer review?  Yes   No   Not clear
*Did the data obtained and the
analysis conducted answer the
research question?  Yes   No   Not clear
Were the measuring instruments
reliable and valid?  Yes   No   Not clear
*Were important extraneous
variables and bias controlled?  Yes   No   Not clear
*If an intervention was tested,
answer the following five questions:  Yes   No   Not clear
1. Were participants randomly
assigned to groups and were
the two groups similar at the
start (before the intervention)?  Yes   No   Not clear
2. Were the interventions well
defined and consistently
delivered?  Yes   No   Not clear
3. Were the groups treated
equally other than the
difference in interventions?  Yes   No   Not clear
4. If no difference was found, was
the sample size large enough
to detect a difference if one existed?  Yes   No   Not clear
5. If a difference was found, are
you confident it was due to the
intervention?  Yes   No   Not clear
Are the findings consistent with
findings from other studies?  Yes   Some   No   Not clear
ARE THE FINDINGS CREDIBLE?  Yes All   Yes Some   No
Clinical Significance
Note any difference in means, r2s, or measures of clinical effects (ABI, NNT, RR, OR)
*Is the target population clearly
described?  Yes   No   Not clear
*Is the frequency, association, or
treatment effect impressive enough
for you to be confident that the finding
would make a clinical difference if used
as the basis for care?  Yes   No   Not clear
ARE THE FINDINGS
CLINICALLY SIGNIFICANT?  Yes All   Yes Some   No
* = Important criteria
Comments
___________________________________________________________________________
___________________________________________________________________________
___________________________________________________________________________

Leveraging Machine Learning and Smart Devices for Diabetes Management
Introduction
Diabetes is a chronic condition affecting millions worldwide that requires ongoing self-management. Advances in technology have the potential to transform diabetes care through personalized interventions that support patients. Machine learning and smart devices show promise for improving outcomes, but their effectiveness needs rigorous evaluation. This article examines the findings of a recent systematic review exploring the use of these technologies for diabetes management.
Methods

Makroum et al. (2022) conducted a systematic search of literature from 2011 to 2021 to identify studies applying machine learning techniques with smart devices to aid diabetes care. A total of 19 19 19 eligible studies were critically appraised based on research methodology, credibility of findings, and clinical significance. Various machine learning approaches like support vector machines and artificial neural networks were explored. Outcomes of interest included glycemic control, self