Supported App Versions, Operating Systems and Devices

The pivotal clinical study was undertaken in five U.S. nursing home (4 in Iowa and 1 in New York) enrolled 104 subjects, 103 were included in the Primary Efficacy Analysis.

Subjects enrolled in the study were required to meet all of the following criteria:

1)      Have been diagnosed with moderate-to-severe dementia on the basis of the score obtained with one of the following tools:

a.   Mini-Mental State Examination (MMSE) score < 19 OR

b.   MDS Cognitive Performance Scale (CPS) score > 3 OR

c.   Another cognitive assessment tool routinely used within the aged care facility allowing the

identification of subjects with moderate-to-severe dementia

2)      They are unable to reliably self-report pain as determined by the caregiver.

3)      Have been living in the aged care home for at least 30 days prior to the day of screening.

4)      Must have had an informed consent signed by the subject's legally authorized representative.

Subjects were not enrolled (i.e. excluded) from the study if they met any of the following criteria:

1)     Subject is unable to partially or completely exhibit any facial expression (e.g. as a result of a facial palsy, facial injuries or other pathologies).

2)     The treating physician determines that it is inappropriate to assess the subject for pain.

3)     Subject is currently receiving or has received in the last 30 days prior to screening an investigational product and/or participated in another clinical trial.

Weekly pain ratings of enrolled subjects were performed concurrently by two independent raters using PainChek^® Adult and the Abbey Pain Scale (APS). Pain assessments were completed twice (at rest and post movement) at each session. In addition, the study coordinator documented changes in clinical status of the resident between each assessment visit. Raters were blinded to each other’s assessment, to prior pain assessments on the same subject, and to subject’s clinical status as assessed by the study coordinator. Each subject was planned to be assessed up to 10 times for a maximum study duration of 16 weeks per subject.

The primary endpoints in this study were agreement rates between the PainChek^® output and APS raters for each pain intensity level. The agreement rate was determined based on a comparison between the categorical output of the PainChek^® Adult and the APS pain category measured by independent raters and compared against a predetermined performance goal for each pain category.

The demographic data of the 103 Residents included in the Primary Efficacy Analysis Set are shown in Table 1.

^{Table 1: Summary of Subjects Demographic Characteristics (Primary Efficacy Analysis Set)}

High Agreement with Abbey Pain Scale: PainChek^® demonstrated a strong agreement (point estimates of the rate of agreement > 70%) with the widely-used Abbey Pain Scale across all pain intensities for assessment-level analyses, underscoring its accuracy as a pain assessment tool (Table 2).

^{Table 2: Frequency Distribution of Abbey Pain Categories versus PainChek® Pain Categories at the Assessment Level (Primary Efficacy Analysis Set)}

Study success criteria were based on agreement rates between APS and PainChek^® Adult for the lower 95% confidence interval for “No Pain” and “Mild Pain” categories and based on the point estimate for “Moderate Pain” and “Severe Pain” categories. Table 3 presents outcomes of the primary effectiveness endpoint for agreement rates calculated at the subject level using the methodology proposed by Emir et al. (Statistics in Medicine, 1998; 17:2563-78.). In this analysis:

^{Table 3: Agreement Rates Between Abbey Pain Scale (APS) and PainChek by Pain Category at the Individual Subject Level (Primary Efficacy Analysis Set)}

^{[1] The rate of agreement and the confidence intervals were constructed following the methodology inEmir et. al. (1998).}

^{[2] For “Moderate Pain” and “Severe Pain” categories the confidence interval is presented as supportive information. The success for these categories is determined based on the point estimate of agreement rate.}

Validated Concurrent Accuracy: Secondary analyses using Deming regression (See Table 4) and Bland-Altman plots (See Figures 1-3) further validated the concurrent accuracy of PainChek^®, showing it aligns closely with the Abbey Pain Scale in assessing pain.

^{Table 4: Mean Difference and Limits of Agreement (Efficacy Analysis Set)}

^{Figure 1: Bland-Altman Plot for APS versus PainChek Z-Scores for All Assessments (Primary Efficacy Analysis Set)}

^{Figure 2: Bland-Altman Plot for APS versus PainChek Z-Scores for Assessments “At Rest” (Primary Efficacy Analysis Set)}

^{Figure 3: Bland-Altman Plot for APS versus PainChek Z-Scores for Assessments “Post Movement” (Primary Efficacy Analysis Set)}

Reliable Test-Retest Consistency: The tool exhibited strong reliability in repeated assessments, ach- ieving an Intraclass Correlation Coefficient (ICC) of 0.73 with was higher than that of Abbey Pain Scale (0.59) as shown in Table 5, affirming its consistency in providing accurate pain evaluations.

^{Table 5: Intraclass Correlation Coefficient (Test-Retest Analysis Set)}

Table 6 presents the frequency distribution of PainChek^® Adult assessment 1 versus PainChek^® Adult assessment 2 for the test-retest analysis set. The overall PainChek^® Adult rate of agreement was 66.7%. Considering Assessment 1 as “Reference” and Assessment 2 as “Test”, the rates of agreement of PainChek^® Adult to itself were 82%, 56%, 52.9% and 25% for “No Pain,” “Mild Pain,” “Moderate Pain,” and “Severe Pain,” respectively. Considering Assessment 2 as “Reference” and Assessment 1 as “Test,” the rates of agreement of PainChek^® Adult to itself were 76%, 66%, 39% and 25% for “No Pain,” “Mild Pain,” “Moderate Pain,” and “Severe Pain,” respectively.

^{Table 6: PainChek: Test-Retest Agreement Analysis for Pain Categories (Test-Retest Analysis Set)}

Further supporting the Test-Retest Reliability of PainChek^® are the results of Deming Regression analysis. This is demonstrated in the scatter plot below (Figure 4) along with the regression line and the unity line. As can be seen from the graphical representation, the regression line is sufficiently close to the unity line and support PCK repeatability.

^{Figure 4: Scatter Plot of PainChek Test-Retest Results Along with Deming Regression (Test-Retest Analysis Set).}

^{All Assessments, All Patients, Replicate #2 Serves as Reference}

^{Note: reference lines divide the numeric pain scores into pain categories; No Pain = 0-6; Mild Pain = 7-11; Moderate Pain = 12-15; Severe Pain = 16+}

Broad Applicability: PainChek^® performed reliably across diverse factors, with results unaffected by age, gender, race, ethnicity, chronic pain history, use of pain medications, or device platform (iOS or Android), making it highly versatile in various clinical settings.

These findings reinforce that PainChek^® Adult aligns with scientific literature as a valid, reliable, and responsive tool for pain assessment in dementia care by healthcare professionals in the nursing home setting.

It is important to note that PainChek^® Adult is an observational pain assessment tool intended to assist trained healthcare professionals in assessing pain in non-verbal patients. As such, PainChek^® Adult is not a standalone diagnostic tool and its results should be interpreted by qualified healthcare professionals within the context of their broader clinical knowledge and examination of the Resident.

In addition, PainChek^® users should be aware that although PainChek^® clinical study data support the accuracy of the device for identifying No Pain, Mild Pain, and Moderate Pain, a low number of residents were identified as suffering from Severe Pain. For this reason, there is uncertainty in the ability of the device to assess Severe Pain, given the small sample size.

Furthermore, when interpreting pain assessment results, users should be aware that the lower bounds of the 95% confidence intervals (CI) for performance across all pain severity categories, particularly Moderate and Severe Pain, indicate increased uncertainty. This uncertainty may be due to the small number of assessments in these categories and the limited test-retest reliability of the clinical comparator used (the Abbey Pain Scale). These factors may have contributed to the observed lower confidence bounds.

It is also important to note that PainChek^® clinical data were evaluated to determine whether facial expressions associated with pain are consistent across racial groups. Although the study population consisted of a less diverse population than that of the United States, the study findings and evidence from published research support the conclusion that core facial expressions of pain are consistent across racial and ethnic groups.

PainChek^® is not a substitute for the clinical judgment or assessment by a qualified healthcare professional, who may decide that PainChek^® is not appropriate for some Residents.