Reliability, Validity and Responsiveness of the Slovenian Version of the Patient Evaluation Measure (PEM-Slo) in Patients with Wrist and Hand Disorders
, and | Oct 04, 2023
Article Category: Original scientific article
Published Online: Oct 04, 2023
Page range: 198 - 206
Received: Jun 06, 2023
Accepted: Sep 22, 2023
DOI: https://doi.org/10.2478/sjph-2023-0028
Keywords
© 2023 Vida BOJNEC et al., published by Sciendo
This work is licensed under the Creative Commons Attribution-NonCommercial-NoDerivatives 3.0 License.
The hand is of paramount importance for humans to interact with their environment. Any impairment of hand function can have an enormous impact on one’s wellbeing. When assessing patients with hand disorders in the clinical and research environment, objective measures such as range of motion, joint stability and grip strength are correlated with subjective patient reported outcome measures (1, 2). The latter give us information about functioning and quality of life from the patient’s perspective in their own specific living and working environment, which cannot be obtained with mere objective measurements (2).
The selection of the patient reported outcome measure for clinical and research work should be based on its simple use regarding clarity and time to answer the questions (3). The results obtained should be comparable to the results of other studies and not influenced by differences in the socio-cultural environment (4). When the patient reported measure is used in an environment where a language other than its original version is spoken, the translation and cross-cultural adaptation must follow the uniform principle for translation and cross-cultural adaptation (4). According to the COnsensus based Standards for the selection of health Measurement INstruments (COSMIN) group, the patient reported outcome measure should have good psychometric properties in terms of being reliable, valid and responsive (5). It is reliable when it is free of measurement errors, valid when it measures what it is intended to measure, and responsive to detect change over time in the construct it measures (5).
The Patient Evaluation Measure (PEM) was first introduced in 1995 to be used in patients with hand surgery (2), followed by its application in patients with diverse hand disorders, i.e. after carpal tunnel surgery (6, 7), distal radius fracture (8), scaphoid fracture (9) and in patients with miscellaneous hand pathology (10). It is a self-administered, region-specific measure for patients with wrist and hand disorders and consists of 18 items divided into three sections. The first section (Part One: Treatment) has five items evaluating the treatment process and is not included in the final score. Section two (Part Two: How your hand is now) is focused on hand and wrist impairments and contains ten items, of which three are about the symptoms (e.g. feeling and pain), five about hand function (e.g. movement and grip), and two items about the impact the hand impairment has on the patient (e.g. daily activities and work). Dias et al. have added one item to the second section, asking about the duration of the pain in the hand (9). The third section (Part Three: Overall assessment) has only three items asking about general satisfaction with the treatment and hand status (2). The PEM is easy to understand, requires little time for completion and can easily be analysed, which makes it more appropriate for use in daily clinical practice (6, 8, 10) compared to e.g. the Michigan Hand Questionnaire. PEM unanimously assesses only one hand which is advantageous in comparison to DASH and QuickDASH.
The aim of this study was to translate and cross-culturally adapt PEM into Slovenian and further evaluate its psychometric properties.
The study was a single-centre observational prospective study conducted from July 2020 to March 2021. Ethical approval was obtained from the UMC Maribor Ethics Committee on May 11, 2020 (UKC-MB-KME-20/20). Written informed consent was obtained from all the participants prior to inclusion in the study.
The study participants were outpatients with hand and wrist disorders aged 18 years and above, all native Slovenian speakers. The exclusion criteria were 1) not being able to answer the questionnaire due to cognitive impairment, 2) bilateral disorders of wrists or hands, 3) rheumatologic or neurologic disease affecting the upper extremities.
PEM is a self-administered, region-specific patient-reported outcome measure for patients with wrist and hand disorders and consists of 18 items divided into three sections. Part One is not included in the final score (2). The items of PEM are scored on a scale from 1 to 7, where 1 represents the best and 7 the worst score (2, 11). The previous validation study of PEM calculated the PEM score by summing the values of each item in Part Two and Part Three, and expressed the sum of the two as a percentage of the maximum score (9). Since no specific instructions for the calculation were given, we developed a formula to express a score on a scale as percentage points from 0–100, a higher score representing a worse functional outcome. The formula used to calculate the PEM score is as follows: ((calculated sum–lowest possible sum)/range) ×100 i.e. ((calculated sum–13)/78) ×100.
The Disabilities of the Arm, Shoulder and Hand questionnaire (DASH) is a region-specific patient-reported outcome measure for patients with upper extremity impairments (12). It consists of 30 items scored on a five-point scale where 1 represents the best and 5 the worst score. The official Slovenian translation of DASH was used (13).
The EQ-5D-5L is a generic patient-reported outcome measure assessing general health related quality of life (HRQoL) (14). It consists of five dimensions rated on a five-level descriptive system. Each combination of answers has an index-based value assigned ranging from -0.495 to 1 in a given crosswalk value-set for Slovenia (15). An official Slovenian translation of EQ-5D-5L was used in the study.
The Global Rating of Change (GROC) scale used in our study is an 11-point scale allowing patients to rate their improvement. It was used as the external anchor to dichotomise the patients into improvers and non-improvers. The question for the GROC scale was formulated as follows: “How is your hand today compared to the day when you first completed the questionnaires?” The possible answers were 0 for no change, -1 to -5 as deterioration and +1 to +5 as improvement. The minimal clinically important change for an 11-point GROC scale is 2 points (16). The arbitrary value for a meaningful improvement was set at 3 points and above to point out the meaningful improvement surpassing the 2 points that represent the minimal clinically important difference (MCID). According to der Roer et al. a change in scores bellow the MCID value should be considered as irrelevant to patient improvement, whereas changes that fall within the range of MCID can be influenced by other factors (such as patient satisfaction with therapy or return to work) and may be false since they may lie within the measurement error (17).
Grip strength was measured with a hydraulic Jamar® hand dynamometer (Sammons Preston Rolyan Inc., USA) following a standardised procedure (18). The measurements were taken three times with each hand and the mean value for each hand was calculated. The values given are expressed as the percentage of the grip strength value of the contralateral unaffected side.
The first stage of the study was the translation of PEM following the guidelines proposed by Beaton et al. (4). The second part of the study was cognitive interviews with twelve patients with hand or wrist disorders.
The third stage was the assessment of the psychometric properties of PEM-Slo. According to COSMIN guidelines, a sample size of 50 participants is the objective for a study to have good quality status when assessing psychometric properties (19). Fifty-one native Slovenian-speaking patients with unilateral wrist and hand disorders treated at our rehabilitation institute were included in the study. At their first outpatient appointment (T1), the participants were asked to complete the three patient-reported outcome measures (PEM, DASH and EQ-5D-5L). They completed the PEM for the second time (T2) after two to four days. All three patient-reported outcome measures were completed again at the end of the outpatient rehabilitation programme (T3). Grip strength was measured twice, at T1 and T3 time points. At T3, participants also assessed their level of improvement on the GROC scale.
The original English version of PEM was independently translated into Slovenian by two native speakers: a physical and rehabilitation medicine (PRM) specialist and a professional translator. Two bilingual native English speakers without medical background and without being familiar with the original English version performed the back-translation. The two back-translated versions were compared with each other and with the original version, where no alteration in any question connotation was observed. At the expert group meeting consisting of the professional translator, linguistic expert for Slovenian and two PRM specialists, the final PEM-Slo version was reached unanimously.
Cognitive interviews were performed on a group of 12 native Slovenian-speaking patients (6 men) with an average age of 52 years (range 34–84) with hand and wrist disorders treated at our rehabilitation institute. The interviews were conducted by one of the authors (ZKR) who provided guidance when answering the PEM-Slo question by question, ensuring that any uncertainties of understanding the questions was reported and discussed. Only some minor uncertainties were raised. Those were regarding the questions in the first part, where it was not clear to the participants which doctor the questions were about since they were treated by a trauma surgeon before visiting a PRM specialist. Another ambiguity was the time point in Part Two, when participants were not certain what the word “now” signifies. They raised questions such as: “Do you mean now, at this moment, or today, or in the last few days?” Since the described uncertainties are not related to possible translational flaws or cross-cultural differences no change to the questionnaire was made.
The reliability is calculated by Cronbach’s α, where values above 0.7 should be reached, indicating that items are adequately correlated and measure the same construct (5,20). The PEM completed at T1, T2 and T3 was used in this analysis. Another reliability measure is expressed with ICC measuring the stability level of the patient-reported outcome measure over time at repeated measurements when no true health status change is expected (5). ICC was used for test-retest reliability purposes: two-way mixed effects, consistency, single rater/measurement model selection (21).
The PEM completed at T1 and T2 in the interval of 2 to 4 days when no true health status change is expected were used in this analysis, as well as in calculating the measurement error expressed with standard error of measurement (SEM), defined as a change in the patient’s score occurring by chance (5). The SEM was used to calculate the minimal detectable change (MDC) (22).
Construct validity of PEM was assessed by calculating the correlations among PEM, DASH, EQ-5D-5L and grip strength. We hypothesised the correlations to be high between PEM and DASH, PEM and some items of EQ-5D-5L (self-care, usual activities, pain), and PEM and grip strength since they measure similar constructs. We expected to find a low or no correlation between PEM and some EQ-5D-5L items, which measure different constructs (i.e., mobility, anxiety/depression). Values obtained at T1 were used in the analysis. The correlations are considered high when the correlation coefficient is >0.6, moderate when r=0.3–0.6 and low when r<0.3 (23).
To evaluate the ability of the PEM, DASH and EQ-5D-5L to detect clinical change, the standardised response mean (SRM) was calculated from the mean score improvement divided by standard deviation (SD) of score improvement. The SRM is considered large when above 0.8, moderate when the value is between 0.5–0.8, and small when less than 0.5 (24). The T1 and T3 values were used in this analysis. Effect size (ES) with Cohen’s d was calculated for all the patient-reported outcome measures in the study using T1 and T3 values. The ES is interpreted as small when d=0.2, moderate when d=0.5, and strong when d=0.8 (25).
The anchor-based approach was used for the MCID evaluation with the receiver operating characteristic (ROC) method using the GROC scale as the external anchor to dichotomise patients. The discriminative power of PEM to differentiate improvers from non-improvers was estimated with the area under curve (AUC), where values less than 0.5, 0.5 to 0.7, 0.7 to 0.8, 0.8 to 0.9, and above 0.9 denote no, poor, acceptable, excellent and outstanding discrimination, respectively (26).
Floor or ceiling effects are present if more than 15% of the population achieves the lowest or highest possible score (22). Values of PEM at T1 and T3 were used in this analysis.
Simple descriptive statistics was used for the analysis of demographic and clinical patients’ data. The IBM SPSS 26 was used for data analysis (IBM Corp., Armonk, NY). The data distribution was analysed using the Shapiro-Wilk test. The Wilcoxon signed-rank test was used to compare the initial and final values (level of significance, p<0.05). Internal consistency was calculated with Cronbach’s α. ICC with 95%CI was used for calculating test-retest reliability. SEM was calculated from the square root of variance between measurements and the error variance of the ICC. MDC with a confidence level of 95% was calculated using the following formula: MDC=SEMx1.96x√2 (22). Spearman’s correlation coefficient was used to assess correlations among variables. ROC curve was plotted for PEM and its cut-off point was determined by the maximum value of Youden’s index: sensitivity + specificity –1 (27). The AUC was estimated for the accuracy of PEM to distinguish improvers from non-improvers. ES and SRM were calculated for PEM, DASH and EQ-5D-5L by comparing values at T1 and T3.
The total number of patients in the study was 51 (25 men), and the mean age was 53 years. The most common type of injury was distal radius fracture. The average timespan between T1 and T2, and T1 and T3 was 2.3 and 22.2 days, respectively. The patients received on average 10 rehabilitation sessions. The demographic and clinical data is presented in Table 1.
Demographic and clinical data of patients (n=51), mean number of rehabilitation sessions, and timespan between assessments.
| Patients Rehabilitation sessions Timespan between assessments | Number Number Days |
|---|---|
| 51 | |
| 53.3 (18-78) | |
| 25 (49) | |
| 25 (48.1) | |
| 2 (3.8) | |
| 3 (5.8) | |
| 3 (5.8) | |
| 7 (13.5) | |
| 5 (9.6) | |
| 2 (3.8) | |
| 4 (7.7) | |
| 29 (56.9) | |
| 16 (31.4) | |
| 10.2 (3-23) | |
| 2.3 (2-4) | |
| 22.2 (10-80) |
Legend: T1 - initial assessment at the beginning of the rehabilitation programme; T2 - second assessment; T3 - final assessment at the conclusion of the rehabilitation programme.
Internal consistency for PEM expressed with Cronbach’s α at T1 was 0.932, at T2 0.938, and at T3 0.953. The ICC values with 95%CI for individual items and the overall PEM score are given in Table 2. The SEM was 6.497 and used to determine the MDC95, which was 18.01.
Test-retest reliability of PEM-Part Two, PEM-Part Three, and PEM total expressed with intraclass correlation coefficient (ICC) with 95% confidence interval (95%CI).
| PEM item/question (Q) | ICC (95%CI) |
|---|---|
| Q1 | 0.599 (0.228-0.774) |
| Q2 | 0.791 (0.630-0.882) |
| Q3 | 0.824 (0.687-0.901) |
| Q4 | 0.789 (0.626-0.881) |
| Q5 | 0.798 (0.642-0.886) |
| Q6 | 0.684 (0.440-0.882) |
| Q7 | 0.848 (0.730-0.914) |
| Q8 | 0.871 (0.770-0.927) |
| Q9 | 0.714 (0.493-0.839) |
| Q10 | 0.825 (0.690-0.901) |
| Q1 | 0.702 (0.471-0.832) |
| Q2 | 0.801 (0.647-0.888) |
| Q3 | 0.740 (0.539-0.853) |
| 0.874 (0.777-0.929) |
Legend: Q - question, item; PEM - Patient Evaluation Measure; ICC - intraclass correlation coefficient; 95%CI - 95% confidence interval.
A high correlation with r=0.757 was proved only between PEM and DASH, moderate correlations with r=0.415–0.6 between PEM and grip strength, and EQ-5D-5L–Self-care, EQ-5D-5L–Usual activities, EQ-5D-5L–Pain, EQ-5D-5L– Anxiety/Depression and EQ-5D-5L¬–total. No correlation was found between PEM and EQ-5D-5L–Mobility (r=0.185 was not statistically significant) (Table 3).
Correlations of different outcome measures with PEM at baseline (T1,
| PEM (r) | ||
|---|---|---|
| Grip strength | 0.415 | 0.003 |
| DASH | 0.757 | <0.001 |
| EQ-5D-5L | 0.600 | <0.001 |
| EQ-5D-5L-Mobility | 0.185 | 0.193 |
| EQ-5D-5L-Self-care | 0.547 | <0.001 |
| EQ-5D-5L-Usual activities | 0.505 | <0.001 |
| EQ-5D-5L-Pain | 0.575 | <0.001 |
| EQ-5D-5L-Anxiety/Depression | 0.417 | 0.002 |
Legend: PEM - Patient Evaluation Measure; DASH - Disability of Arm, Shoulder and Hand Questionnaire; EQ-5D-5L - EuroQol questionnaire; r - Spearman’s correlation coefficient.
All the observed changes between baseline and follow-up measurements were statistically significant (Table 4). PEM and DASH were highly responsive with ES and SRM 1.25 and 1.42, and 0.85 and 1.08, respectively. EQ-5D-5L showed only moderate responsiveness with ES 0.66 and SRM 0.78.
Total scores of outcome measures at baseline (T1), second assessment (T2), and follow-up (T3) for all patients (n=51), observed change (T3)-(T1), standardised response mean (SRM) and effect size (ES).
| Outcome measure | Baseline (T1) mean (SD) (range) | Two to four days after T1 (T2) mean (SD) (range) | Follow-up (T3) mean (SD) (range) | Observed change (T3) - (T1) * mean (SD) (range) | ES | SRM |
|---|---|---|---|---|---|---|
| 44.1 (18.7) | 37 (17.9) | 22 (16.6) | - 21 (14.8) | 1.25 | 1.42 | |
| (3.9-82.1) | (3.9-69.2) | (0-73.1) | (-3.8- -68) | |||
| 41.6 (23.5) | / | 23.4 (18.9) | -18.8 (17.4) | 0.85 | 1.08 | |
| (5-91) | (0-71.6) | (-55.2- -21.6) | ||||
| 0.7 (0.15) | / | 0.799 (0.15) | 0.097 (0.125) | 0.66 | 0.78 | |
| (0.195-1) | (0.522-1) | (-0.193-0.506) | ||||
| 40.6 (26.4) | / | 61.2 (26.2) | 19.7 (19.9) | 0.81 | 0.99 | |
| (7-99) | (9-108) | (-10-70) |
Legend: PEM - Patient Evaluation Measure; DASH - Disability of Arm, Shoulder and Hand questionnaire; EQ-5D-5L - EuroQol questionnaire; SD - standard deviation; SRM - standardised response mean; ES - effect size;
* Wilcoxon signed-rank test with level of significance p<0.001.
Twenty-six patients with GROC score 3 or above represented a cohort with a meaningful improvement whereas 25 of them with a GROC score of 2 or less fell into the cohort without meaningful improvement. The two cohorts were used to plot the ROC curve, which is shown in Figure 1. The cut-off point was determined with Youden’s index representing the MCID of PEM-Slo in our study, which was 17.31 with a 76.2% sensitivity and 80.8% specificity. AUC was 0.869 (95%CI 0.769–0.969).
Figure 1.
Receiver operating characteristics (ROC) curve for PEM with an 80.8% specificity and 76.2% sensitivity, area under curve (AUC) 0.869 (95%CI 0.769-0.969), cut-off value representing minimal clinically important difference (MCID) of 17.31.
No maximal or minimal values of PEM were attained at T1, whereas only one minimal and no maximal score was observed at T3. This represents 1.9% of the total sample reaching minimal (best) value at T3.
The psychometric analysis of this study shows the PEM-Slo to be a reliable, valid, and responsive patient-reported outcome measure for patients with unilateral hand and wrist disorders. The reliability assessment with Cronbach’s α above 0.9 at all time points shows the PEM-Slo to have an excellent internal consistency with highly interrelated items of PEM (20). Hobby et al. obtained the value of Cronbach’s α of 0.94 in patients after surgical decompression of the carpal tunnel (CT) (6). This is similar to the findings in the non-specified cohort of patients comparing three patient-reported outcome measures (0.91 for Part Two and 0.88 for PEM Part Two and Part Three) (11), whereas the same value of Cronbach’s α as in our study (0.932) for PEM total (Part Two and Part Three) was reported in patients with scaphoid fracture (9). Cronbach’s α in patients with distal radius fracture using only Part Two of the questionnaire was 0.94 (8) and in patients after surgical CT decompression was above 0.9 at all time points of the study (7).
Test-retest reliability of PEM-Slo was in the range of good stability level of the patient-reported outcome measure over a time span of 2.3 days on average (range 2-4 days) which is too short for a true health status change in hand and wrist disorders to occur with the ICC value of PEM being 0.874. Our results are in line with the study where the □ value of 0.83 for PEM total was obtained (11). The low ICC for item 1 in the PEM score may indicate the uncertainty of participants in interpreting the term “feeling in the hand”. A similar finding was described by Dias et al., where the word “feeling” in the item “The feeling in my hand is normal…completely absent” may connote different meanings besides feeling of touch, such as “feeling of stiffness, feeling swollen or feeling not right” (9). They have proposed a more precise explanation of the feeling in item 1, such as “feeling of touch”, which could clarify the focus of the question on the assessment of sensation. However, item 1 in the original version of PEM as well as in PEM-Slo remained unchanged.
The construct validity of PEM in our study was established using hypothesis testing with other outcome measures for hand function. The convergent validity was proved with a high correlation of PEM and DASH, and moderate correlation of PEM and grip strength, EQ-5D-5L and parts of EQ-5D-5L, which can be influenced by impaired hand function (self-care, usual activities, pain, anxiety/depression). The divergent validity was demonstrated with no correlation of PEM with mobility dimension of EQ-5D-5L, the latter assessing problems in walking about where the hand function in our cohort of patients is insignificant. Correlation between PEM and DASH was stronger than the correlation of both DASH and PEM with grip strength, since both are region-specific instruments assessing a similar construct. Similar findings were reported in other studies with strong correlation of PEM and DASH where the values of correlation coefficients were 0.73 (8) and 0.82 (10). Moderate correlations were also found among wrist and finger range of motion and PEM (0.57 and 0.41, respectively) (10). The range of motion was not used in the analysis in our study since it could not be directly compared among participants due to the miscellaneous cohort of patients.
The PEM was established as the most responsive questionnaire to detect change among the three patient-reported outcome measures in our study. This emphasizes the importance of using a region-specific or disease-specific questionnaire. EQ-5D-5L as a general HRQol measure showed only moderate responsiveness for patients with hand and wrist impairments, whereas PEM and DASH were highly responsive, with PEM’s SRM and ES values surpassing those of DASH. A better responsiveness of PEM in comparison to DASH was also proved in CT syndrome patients (ES for PEM 0.97 and ES for DASH 0.49) (6) and in patients with scaphoid fracture, where the values of ES and SRM for PEM were similar to ours (SRM -1.46 and -1.47, and ES -1.12 and -1.1 at two different time points) (9).
When interpreting the scores of the outcome measures it is important to know the change in score that may occur by chance due to measurement error. The MDC95 for PEM-Slo in our study was 18.01. This represents a lower limit above which the observed change in a scale can be interpreted as a true change. The MCID of a scale corresponds to the value where the observed change represents the meaningful improvement in a patient. The MCID in our study calculated with the ROC curve method was 17.31 with the AUC value of 0.869 indicative of excellent discriminative power. The MDC95 and MCID values for PEM-Slo in our cohort of patients are very close to each other, with MDC95 even slightly surpassing the MCID value. The MDC95 was calculated with the distribution-based method using SEM and the calculations were made on an entire sample (n=51) without using the external anchor to dichotomise patients into improvers and non-improvers. The values obtained with anchor-based methods are more reliable and should be used to establish the MCID in outcome measures (28). However, since both values in our study are very close (MDC95=18.01, MCID=17.31) we propose the change in PEM score above 18 to be considered as a clinically important improvement. To the best of our knowledge there is only one conference paper about the interpretability of PEM reporting the MCID=3 in patients with Dupuytren’s contracture using the ROC (AUC) method (29). However, no direct comparison to this study can be made since it is important to bear in mind that the MCID can vary among different cohorts of patients (30) and methods used to estimate its value (28).
The advantage of PEM over other region-specific patient-reported outcome measures such as DASH or Patient Rated Wrist Evaluation (PRWE) is its unambiguous focus of questions on the impaired hand. DASH evaluates the capacity to carry out a task irrespective of which hand is used to perform the task (12). The PRWE is clearly focused on the affected hand (31), yet some questions are assessing the function of the dominant hand (e. g. “cutting meat using my affected hand” or “using bathroom tissue with the affected hand”) where patients can have difficulties in choosing the appropriate answer if they never use the affected hand for such tasks. Items of PEM in Part Two from 1-6 and 9-10, and items 2 and 3 in Part Three are focused on the assessment of one hand whereas items 7 and 8 about function in Part Two and item 1 in Part Three could assess both hands simultaneously. It may be reasonable to assess separately each side with bilateral disorders, however the items that assess function would probably have lower values with bilateral hand and wrist problems. It is therefore reasonable to limit the use of PEM in unilateral wrist and hand disorders. We believe that none of the items in PEM is biased by the handedness of a patient.
The use of PEM in literature is, in our opinion, inconsistent. There are only numbers reported without any clear explanation whether this is the calculated sum of scores, or the number represents the score on a scale from 0–100 (32). Sometimes only Part Two of PEM is used in the analysis instead of Part Two and Part Three as suggested by the developers (8, 33). Our study clarifies the use of the scale as suggested by the developers. It appeals to a more consistent use of PEM using Part Two and Part Three in the calculation of the final score with the formula that transforms the result on a scale from 0-100, which represents a more clear-cut presentation of the results. The major shortcoming of our study is a relatively small sample size. The total number of 51 patients in our study falls in the lower limit of necessary subjects according to Terwee et al. (19) when assessing the construct validity of the instrument in relation to other measures. A larger sample size would also be needed to check for the unidimensional structure of the instrument with confirmatory factor analysis, where a sample size of 150 to 1,000 subjects is recommended in the literature (34). In the previous validation studies of PEM the sample sizes were in the range from 32 to 200 patients (number of participants were as follows: 32 (6), 35 (11), 50 (7), 80 (9), 100 (10), 200 (8)). According to Bacchetti, who recommends choosing a sample size that was used “in the past for similar or analogous studies” (35) we aimed for a sample size of 100 patients. The study was conducted in the time of COVID where the number of patients in our rehabilitation ward was reduced to a minimum (36), and the study time was limited which resulted in a smaller sample size than we aimed for, with the total of 51 patients. Another shortcoming of our study related again to the small sample size is calculation of the internal consistency expressed with Cronbach’s alpha coefficient based only on the assumption of unidimensionality of the scale, since the confirmatory factor analysis to check the structure of the scale could not be performed due to the small sample size. The other shortcoming is the heterogeneous cohort of patients with hand disorders, especially when determining the MCID where the values depend on the characteristics of the examined cohort.
The PEM-Slo is a reliable, valid and highly responsive patient-reported outcome measure for patients with hand and wrist disorders. It has excellent internal consistency and good test-retest reliability. Its construct validity was demonstrated in correlation to DASH, EQ-5D-5L and grip strength. It has no floor or ceiling effect. Our study adds new knowledge about the interpretability of PEM with values of the MDC95 and MCID. It also precisely describes the calculation method for the PEM score, using Part Two and Part Three and expressing it as a value from 0–100. We believe the PEM-Slo is ready to be used in the clinical and research environment. Further research is needed about the interpretability of PEM on more homogenous cohorts where the MDC and MCID values could differ from the ones obtained in our study.