This page provides an independent statistical analysis and educational interpretation of publicly reported results. ClinicalTrials.gov provides the official trial registry record. Numerical trial results on this page are restricted to the results posted in the ClinicalTrials.gov record; calculations added by Clinical Biostats are labelled as such.
1. Trial at a Glance
This page covers the registry results for AI-Powered Fall Risk Prediction in Nursing Care (NCT07000981), a registered phase not stated study. The primary outcome is Fall Risk Categories Based on the Morse Fall Scale.
| Feature | Registry value |
|---|---|
| ClinicalTrials.gov identifier | NCT07000981 |
| Phase | Phase not stated |
| Condition | Accidental Falls, Risk Assessment, Nursing Assessment |
| Allocation | Not Stated |
| Design model | Not Stated |
| Masking | Not Stated |
| Enrollment | 177 |
| Arms in the primary outcome | n/a |
| Lead sponsor | Inonu University |
| Results posted | Yes |
2. Clinical Question
Population
Adults aged 18 years or older who can read and write in Turkish, walk with or without assistance, and voluntarily participate are eligible according to the registry description. The listed exclusions are inadequate ability to speak or understand Turkish, intubation, physical restraint, immobility, and diagnosed cognitive impairment. The record reports enrollment of 177 participants but does not provide their demographic characteristics or the number meeting each eligibility criterion.
Intervention
The study aims to develop a nursing clinical decision-support system that uses machine learning and computer vision for fall-risk prediction. Participants' basic information, including age, height, and weight, is collected using a demographic form. The planned assessment includes walking in a clinical corridor at a comfortable pace while a digital camera-based system records movement; an accelerometer in a pocket records three-axis acceleration during walking. The registry describes the system's intended use, but does not report its predictive performance or feasibility results.
Comparator
The registry describes an initial fall-risk assessment using the Morse Fall Scale, followed by the camera-based walking assessment and accelerometer recording. It does not report a randomized comparator or a separate control arm. Allocation is not stated, and the record cautions that any arms are not a like-for-like control comparison. The Morse scale therefore serves as a described clinical assessment in the study procedures; the registry does not give a comparative analysis between that scale and the proposed system.
Primary question
The study asks whether machine-learning models can serve as valid tools for fall-risk prediction and whether the proposed system is feasible in clinical environments. The named primary outcome is participants' fall-risk categories based on the Morse Fall Scale, assessed on Day 1. That measure describes a clinical risk categorization, not by itself the accuracy of a prediction model. The registry record does not report the category counts, model predictions, validation results, or feasibility findings needed to answer the study questions.
3. Trial Design
The registry describes a study to develop a nursing clinical decision-support system for predicting fall risk with machine learning and computer vision. It is registered as completed, with 177 participants enrolled. The recorded study start date is 2025-06-01 and the primary completion date is 2025-07-01. Inonu University is listed as the sponsor. The registry does not state an allocation method or describe a randomized design, and it cautions that arms are not a like-for-like control comparison. The study should therefore not be interpreted as a randomized treatment comparison on the basis of this record.
The participant procedures start with demographic data collection and an assessment using the Morse Fall Scale. Participants then walk at a comfortable pace in a clinical corridor while a digital camera-based computer-vision system records walking. An accelerometer placed in a pocket records three-axis acceleration during that walk. The registry describes these as elements of the study procedure; it does not state the precise order or timing of additional data-processing and model-validation steps.
Eligibility is described for adults aged 18 years or older who can read and write in Turkish, walk with or without assistance, and are willing to participate. Exclusions include inability to speak or understand Turkish adequately, intubation, physical restraint, immobility, and diagnosed cognitive impairment. The registry record does not report the number screened, reasons for exclusion, recruitment setting details beyond the clinical corridor assessment, or participant characteristics.
Three outcome measures are registered, but only the primary outcome is described in the record. The primary measure is Morse Fall Scale risk categories assessed on Day 1. The record does not provide a detailed analysis plan, results for the primary or other outcomes, or the number of analyses performed. It also does not explain the basis for the enrollment target or describe interim analyses.
4. Participant Flow and Accounting
The registry reports how participants moved through each stage of the study. All counts below are exactly as posted; missing counts are not inferred from other stages.
Before assignment: Research Inclusion Criteria * Being 18 years or older * Being able to read and write Turkish * Being able to walk with or without support Study Exclusion Criteria * Not being able to speak or understand Turkish at a sufficient level * Being diagnosed * Being immobile * Being cognitively inadequate for the study
Overall Study
| Stage | Cohort Group |
|---|---|
| STARTED | 177 |
| COMPLETED | 177 |
| NOT COMPLETED | 0 |
5. Baseline Characteristics
The table reproduces the baseline characteristics posted in the registry, which describe the participants analysed at baseline and help show whether the randomized groups were comparable.
| Characteristic | Unit | Cohort Group |
|---|---|---|
| Participants analysed | 177 | |
| Age, Continuous | years | 52.38 (SD 16.55) |
| Sex: Female, Male: Female | Participants | 87 |
| Sex: Female, Male: Male | Participants | 90 |
| Race and Ethnicity Not Collected | Participants | — |
| Region of Enrollment: Turkey | Participants | 177 |
| BMI | kg/m^2 | 27.21 (SD 5.64) |
6. Primary Endpoint
The registry lists the primary outcome as Fall Risk Categories Based on the Morse Fall Scale, measured over Day 1.
| Element | Registry entry |
|---|---|
| Result type | Count Of Participants |
| Measure of spread | not applicable |
| Unit | Participants |
| Endpoint classification | Binary Endpoint |
The primary outcome is titled “Fall Risk Categories Based on the Morse Fall Scale.” It is recorded as a count of participants and has a Day 1 timeframe. The scale uses six criteria: secondary diagnosis, history of falls, mobilization support, intravenous access or heparin use, gait or transfer, and mental status. The registry describes a score below 25 as low risk, a score between 25 and 50 as medium risk, and a score of 51 and above as high risk. It states that scores can range from 0 to 125.
The Morse scale is used in the study to assign a structured fall-risk category. A standardized measure can make risk assessment more consistent than an informal description, while category counts can show how participants were classified at the assessment. The registry describes the scale as a tool used in hospitals in Turkey and as a systematic way to determine fall risk in clinical settings.
There is an important distinction between the stated primary endpoint and the study's broader aims. The endpoint is a Morse Scale categorization; it is not itself a reported measure of how accurately the computer-vision or machine-learning system predicts falls. The record does not report observed category counts, actual falls, a comparison of predictions with outcomes, or model-performance estimates. Consequently, the endpoint description explains what was assessed, but cannot establish predictive validity or clinical feasibility.
7. Why No Re-analysis Is Shown
8. Statistical Interpretation
The registry identifies the study's purpose, procedures, enrollment, and primary assessment, but it does not report outcome results. In particular, no participant counts in the low-, medium-, or high-risk Morse categories are reported. There are also no reported model predictions, actual fall outcomes, prediction-performance estimates, or feasibility findings. The study status is listed as completed, but completion status is not a substitute for results.
The primary measure is based on the Morse Fall Scale at Day 1. Its described categories depend on the scale score: below 25 is low risk, between 25 and 50 is medium risk, and 51 and above is high risk. Those definitions tell the reader how the assessment classifies risk. Without the observed category counts, however, the record cannot show how risk was distributed among participants. Nor can it show whether the proposed system assigned the same or different risk levels.
The broader study questions concern whether machine-learning models can serve as valid fall-risk prediction tools and whether the system is feasible in clinical environments. Answering a validity question ordinarily requires reported predictions and an appropriate reference outcome or comparator, along with a clear account of how performance was assessed. Answering a feasibility question requires reported implementation or process findings. These results and methods are not reported in the registry record. The description that camera and accelerometer data were collected does not, on its own, establish that the system generated useful predictions or could be used feasibly.
The record lists 177 enrolled participants, but enrollment alone does not establish how many contributed to the primary analysis or whether the data were complete. The analysis population, missing-data handling, and participant flow are not reported. No analysis results or statistical tests are listed. Accordingly, there is no reported estimate or uncertainty interval to interpret, and no basis in the registry record to quantify the precision of any finding.
The registry's adverse-event group lists zero participants at risk, zero participants affected, and zero deaths. With no participants at risk listed, those entries do not establish that no adverse events occurred. They provide no interpretable event rate or safety conclusion. More broadly, the posted information supports a description of the study design and endpoint, not a conclusion about the performance, effect, or safety of the proposed system.
9. Statistical Methods Explained
The following statistical context is general. The registry record does not report the statistical methods used to analyse the study data, so these explanations describe what would ordinarily be needed to interpret the stated questions rather than asserting that any particular analysis was performed.
What does the primary measure count?
The primary outcome is expressed as a count of participants classified using the Morse Fall Scale. The scale description gives three categories: low risk for a score below 25, medium risk for a score between 25 and 50, and high risk for a score of 51 and above. A category count is a descriptive summary. It can show how many assessed participants fall into each class, provided the corresponding counts and analysis denominator are reported. The registry record does not report those counts or a denominator for the primary outcome.
Category counts do not by themselves measure model accuracy. A system that predicts risk would need predictions to be compared with an appropriate reference, and the analysis would need to define what constitutes a correct or useful prediction. For a prediction task, relevant choices may include how categories are assigned, what reference outcome is used, and how performance is summarized. Those choices and results are not reported in the registry record. The fact that Morse Scale categories are collected does not establish that they were used as a validation reference for the model.
What would an uncertainty interval mean?
In general, a confidence interval expresses the uncertainty around an estimate under a stated statistical model and analysis procedure. For category proportions or other summaries, an interval can help show how precisely the sample estimate represents the population quantity being studied. It does not account automatically for every source of bias, such as participant selection or an imperfect reference standard. No estimates or confidence intervals are reported for this study, so there is no numerical precision to assess.
Why does the analysis plan matter?
A statistical analysis plan specifies how outcomes are defined and summarized, which participants are included, how missing observations are handled, and which comparisons or models are used. These decisions affect the interpretation of results. The registry lists three outcome measures and zero analyses, but does not provide an analysis plan or results. It is therefore not possible to identify the planned model, any adjustment variables, inferential tests, or the approach to uncertainty estimation. The record also does not report whether any interim analyses took place.
What does over-dispersion mean in general?
For count outcomes, a statistical model may assume a particular relationship between the average count and its variability. Over-dispersion means that the observed variability is greater than the model assumes. If a model ignores that extra variability, uncertainty can be understated and statistical evidence can appear stronger than it is. Whether over-dispersion is relevant depends on the outcome and model. The registry record does not report a count-regression model or any assessment of dispersion, so no conclusion about this issue can be drawn for the study.
How should a planning calculation be read?
The planning assumptions accompanying the record describe an illustration based on an observed effect, with a two-sided significance level of 0.05, equal allocation between two arms, and a normal approximation. The effect to detect is described as a stated fraction of the observed effect. These are planning assumptions, not a description of the study's own design or proof that the enrollment target was justified by a particular calculation. The record does not report calculated sample-size values or a sample-size justification for the study. Such planning illustrations are not observed power and cannot replace the study's actual outcome estimates.
The assumptions also refer to two arms, although allocation is not stated and the record cautions that arms are not a like-for-like control comparison. Thus, the illustration should not be read as evidence that participants were randomized or that an equal allocation was implemented. A planning calculation only has a meaningful interpretation when its assumed outcome, effect, variance, allocation, and analysis match the intended study question. Those details are not reported as the trial's statistical plan.
What is needed to evaluate prediction?
General methods for prediction assessment compare model outputs with a clearly specified reference and report metrics that fit the intended use. Classification performance may involve agreement, sensitivity, specificity, or other measures, while predicted probabilities can be evaluated for calibration and discrimination. These concepts answer different questions; one measure alone may not capture all aspects of performance. The registry record does not state which measures were used, whether a separate validation set was used, or how data were divided or processed. No performance estimates are reported. It would therefore be inappropriate to infer predictive accuracy from the study's aims or data-collection description.
10. Multiplicity, Missing Data and Analysis Populations
The registry lists three outcome measures, but gives a detailed description only for the primary outcome, Fall Risk Categories Based on the Morse Fall Scale. The other outcome measures and their definitions are not reported in the registry record. The record lists zero analyses, and does not provide results or a statistical analysis plan.
In general, evaluating several outcomes or making repeated comparisons can increase the chance of finding an apparently notable result by chance. A prespecified plan can identify which outcomes are primary, which are secondary, and whether any adjustment or other multiplicity strategy is intended. The registry record does not report how multiplicity was addressed, so no claim about adjustment, hierarchy, or selective reporting can be assessed.
Missing-data handling and analysis populations are also not reported. Although enrollment is listed as 177, the registry does not state how many participants had complete Morse Scale assessments, walking recordings, or accelerometer measurements, or how incomplete data were treated. It does not report exclusions after enrollment or define an analysis population. General statistical practice distinguishes the enrolled group from the participants contributing to a particular analysis; without those details, the denominators behind any results cannot be evaluated. No missing-data methods or sensitivity analyses are reported.
11. What the Result Does, and Does Not, Mean
The record establishes that the study aimed to develop a nursing decision-support system using machine learning and computer vision and describes collection of Morse Scale, walking-camera, and accelerometer data. It also identifies the primary endpoint as Day 1 Morse Fall Scale risk categories and gives the criteria used to define those categories. These details explain the study's intended measurement approach.
They do not establish that the system accurately predicts falls, improves assessment, is feasible in clinical practice, or changes patient outcomes. The registry record does not report model predictions, actual falls, prediction-performance measures, category counts, feasibility results, or comparative results. A completed status and an enrollment count do not fill those evidence gaps. The primary endpoint is a risk classification, not a reported evaluation of how well the proposed system predicts future falls.
The information also does not support conclusions about the representativeness of participants or the reliability of any findings. Eligibility excludes some people, including those who are immobile or have a diagnosed cognitive impairment, and the record does not describe participant characteristics or analysis denominators. Whether the study findings generalize to other clinical settings or patient groups cannot be assessed from the registry entry.
These limitations are about what the registry record reports, not a claim that analyses or results do not exist elsewhere. On this record alone, the appropriate interpretation is descriptive: the study was registered as a completed system-development investigation, but its outcome findings and statistical evaluation are not reported.
12. Why This Trial Matters Statistically
This study raises a useful statistical distinction between collecting information relevant to a prediction system and demonstrating that the system predicts well. A camera-based gait assessment and pocket accelerometer can generate data for modelling, while the Morse Fall Scale provides a structured risk categorization. To evaluate prediction, results must connect the model's outputs to a stated reference and report performance with enough detail to judge uncertainty and limitations. The registry describes the data collection and intended questions, but not that evaluation.
The endpoint also matters because it determines what a result would mean. Morse Fall Scale categories describe an assessment at Day 1. They are not the same as observing whether a participant subsequently falls, and they are not themselves a measure of model accuracy. Keeping these concepts separate prevents an assessment category from being misread as evidence of prediction performance.
For clinical prediction research, transparent reporting of participant flow, outcome definitions, analysis populations, missing data, model development, and validation is important. Those details help readers assess whether apparent performance may reflect the data used to build the model and whether conclusions are precise enough to be informative. The registry does not report these statistical details or results, so the questions it poses remain unanswered in the record.
The entry is therefore statistically informative as a description of intended measurement, but not as an evidentiary account of model performance. The distinction is important when interpreting a completed study: completion indicates a registered status, whereas the strength of statistical evidence depends on reported analyses and results.
13. Limitations
- The posted summary statistics are the only data these results rest on; patient-level data, full protocols and statistical analysis plans are not available.
- Recalculated results are simple, unadjusted calculations from posted summaries. They do not replace the sponsor's prespecified analysis.
The primary endpoint is a Morse Fall Scale category, while the stated study questions concern model validity and feasibility. The record does not explain how the category measure relates to a reference outcome for predicting falls, nor does it report actual falls or model-performance measures. This makes it impossible to assess predictive validity from the registry entry.
The record describes a camera-based walking assessment in a clinical corridor and an accelerometer placed in participants' pockets, but does not report technical details of recording, data quality, feature extraction, or model development and validation. It also does not report whether assessments were repeated or how variability in walking or measurement was handled. These omissions limit reproducibility and interpretation of the proposed prediction approach.
Eligibility criteria exclude people who are immobile, physically restrained, intubated, or have diagnosed cognitive impairment, among others. This may limit the range of people represented by the enrolled group, but the registry does not provide participant characteristics or findings with which to judge that issue. Recruitment details, participant flow, missing observations, and analysis denominators are not reported. The registry also provides no statistical analysis plan or outcome results, and only the primary outcome is described among the three registered outcome measures.
Finally, allocation is not stated, and the record cautions that arms are not a like-for-like control comparison. The study should not be read as a randomized comparison. The registry's adverse-event listing has zero participants at risk, which prevents those entries from supporting an interpretation about event occurrence.
14. Related Tutorials
15. Related Statistical Calculators
16. Sources
- ClinicalTrials.gov: AI-Powered Fall Risk Prediction in Nursing Care — NCT07000981.
- Registry record last updated 2025-12-02; status last verified 2025-11.
- Results first posted 2025-12-02.
- Data retrieved from ClinicalTrials.gov on 2026-11-09.
- Group sizes, means, event counts and sponsor-reported p-values are reproduced as posted. Effect estimates, confidence intervals and p-values not attributed to the sponsor are independently calculated from the posted summary statistics.
Continue through the Clinical Biostats statistical pathway
Use this trial's endpoint as a starting point for deeper study of effect measures, confidence intervals, and sample-size planning.
17. Record Summary
NCT07000981, titled “AI-Powered Fall Risk Prediction in Nursing Care,” is sponsored by Inonu University and is listed as completed. The registry reports 177 participants enrolled, a start date of 2025-06-01, and a primary completion date of 2025-07-01. The study aims to develop a nursing decision-support system using machine learning and computer vision.
Participants were to receive a Morse Fall Scale assessment, complete a camera-based walking assessment in a clinical corridor, and wear a pocket accelerometer during walking. The primary outcome is Day 1 Morse Fall Scale risk categories. The record describes the category thresholds but does not report category counts, model-performance findings, feasibility results, statistical analyses, or results for the other registered outcomes. Allocation and the analysis populations are not reported.