Abstract
This study provides age- and sex-specific norms of the Purdue Pegboard Test (PPT) for a middle-aged and elderly population. The data represent useful reference values when assessing manual dexterity in higher age groups.
Manual dexterity refers to the ability to make coordinated hand and finger movements. It also includes grasping and manipulating objects (Makofske, 2011). Apart from the motor aspects, manual dexterity requires the ability to cognitively plan and execute a task.
Previous studies have suggested this association between motor function and cognitive abilities (Leisman et al., 2016; Quandt et al., 2016). Sensorimotor markers have been shown to predict cognitive decline among patients with mild cognitive impairment and Alzheimer’s disease. An Italian study demonstrated that the speed of manual dexterity relates to longitudinal change on Mini-Mental State Examination scores in a population-based cohort of elderly people (Curreri et al., 2018). A Rotterdam study used the Purdue Pegboard Test (PPT), which is a rapid test of manual dexterity, and showed that it may help to identify people at high risk for neurodegenerative diseases (Darweesh et al., 2017).
The PPT assesses uni- and bimanual finger and hand dexterity. Initially, it was developed in the 1940s as a test for personnel selection (Tiffin & Asher, 1948). It is considered one of the top three assessments of hand dexterity because of its good psychometric characteristics, few confounding variables, and high sensitivity for detecting functional impairment (Lezak et al., 2012). Normative data are particularly available in the young and middle-aged populations (Tiffin, 1968) but are relatively scarce for older adults (65 yr and older; Drag & Bieliauskas, 2010). Personnel selection studies have revealed (Tiffin, 1968; Tiffin & Asher, 1948) that manual dexterity deteriorates with aging. It can also negatively affect activities of daily living and independence (Ranganathan et al., 2001). The magnitude and course of deterioration in PPT performance with age have not yet been fully determined, and little is known about factors that influence test performance in the aging population. We administered the PPT to a large cohort of 1,355 randomly selected elderly community-dwelling participants in the setting of the Austrian Stroke Prevention Study (Schmidt et al., 1994) to determine the influence of demographics and common vascular risk factors on test results and to provide adjusted norms of test performance.
Method
Participants
The Austrian Stroke Prevention Study is a community-based cohort study on the effects of vascular risk factors on brain structure and function in elderly people without a history or signs of stroke and dementia among residents of Graz, Austria (Schmidt et al., 1994). Participants were included in the study if they responded positively to an invitation to undergo an extended medical examination at the University Clinic for Neurology, Medical University of Graz, after being randomly selected from the official registry of residents of the city of Graz, Austria, between 1991 and 1994.
Exclusion criteria were a history of neuropsychiatric disease, including previous cerebrovascular accident and dementia or abnormal findings in the neurological examination. The age range for the sample was 40 to 79 yr. Participants were drawn from two study panels (1991–1994 and 1999–2003). They all underwent a structured clinical interview, a physical and neurological examination, assessment of vascular risk factors, laboratory testing, and extensive neuropsychological testing. The sample included a total of 1,355 individuals (775 [57.2%] women and 580 [42.8%] men), of whom 1,248 (92.1%) were right-hand dominant.
The study was approved by the local ethics committee of the Medical University of Graz for experiments using human participants. Written informed consent was obtained from all study participants.
Demographics and Risk Factors
Education level was determined by years of schooling and the highest grade achieved. Risk factors were determined on the basis of history and measurements at the examination, as previously described (Schmidt et al., 1994). Hypertension was based on a history of hypertension with repeated blood pressure readings ≥140/90 mmHg, treatment of hypertension, or readings at the examination exceeding the limit. Diabetes was coded as present if an individual was being treated for diabetes at the time of examination or if the fasting blood glucose level at the examination exceeded 126 mg/dl. Coronary heart disease was determined according to the Rose questionnaire and electrocardiogram findings (Schmidt et al., 1999). Smoking status was coded as current, former, or never, and participants were considered current smokers if they smoked >10 cigarettes a day. Body mass index (BMI) was calculated in kilograms per square meter. Overweight was defined on the basis of BMI. Men were categorized as overweight if the BMI was ≥27.8; women, if BMI was ≥27.3. The cohort characteristics are shown in Table 1.
Cohort Characteristics
Note. BMI = body mass index; IQR = interquartile range; Mdn = median.
Regression Analyses for the Four Subtests
Note. BMI = body mass index; BP = blood pressure.
Purdue Pegboard Test Administration
The PPT was administered according to Tiffin and Asher’s (1948) guidelines. In brief, study participants were asked to place one cylindrical metal peg into as many of 25 holes in a pegboard as possible in 30 s. The test is performed three times, using the right, the left, and both hands, respectively. A fourth subtest, assembly, which consists of assembling pins, collars, and washers within 60 s and correctly alternating both hands, was also conducted. Only one trial per subtest was allowed. Before each subtest, the participants were allowed to put three pins in a hole and perform three assemblies as a part of practice. The total number of pins inserted with the right hand and with the left hand was recorded. For both hands simultaneously, the total number of pairs of pins was counted and recorded. Finally, the number of parts assembled was recorded as the assembly score. Testing was done by a trained psychologist under constant laboratory conditions.
Data Analysis
The assumption of normal distribution was tested by a Shapiro–Wilk test and by means of a Q–Q plot. To verify the assumption of homogeneity of variance, Levene’s test was used. Continuous, normally distributed data are presented as means and standard deviations and continuous, nonnormally distributed data as medians and interquartile ranges. Categorical data are presented as the number of observations and percentages.
A t test for independent groups was used to compare parametric data, whereas a Mann–Whitney U test or Kruskal–Wallis test was used to compare nonparametric data. A multiple linear regression analysis was used to assess the association among sociodemographics, vascular risk factors, and PPT scores. Predictor variables were standard demographical factors and major vascular risk factors. Before running the multiple linear regression analysis, a multicollinearity analysis showed the variance inflation factor was greater than 10 for years of schooling. Therefore, we decided to keep only the highest grade achieved as the demographic factor in the regression analysis.
Normative data are presented stratified by significant determinants of the PPT results. Normative statistics included the mean; 95% confidence interval; standard deviation; and 25th, 50th, and 75th percentile cutoff values.
All statistical analyses were performed using IBM Statistics (Version 26.0). A two-sided p < .05 was considered to be statistically significant.
Results
The association among demographics, vascular risk factors, and PPT performance is shown in Table 2. For all PPT subtests, higher age, male sex, and history of diabetes were consistently associated with worse test results. Educational level, smoking, and BMI were not consistently related to test performance, although significant associations existed for subtests of the PPT.
Age, sex, and diabetes together explained between 25.0% and 31.5% of the variance of the PPT total and subtest performance, with diabetes explaining only between 0.7% and 1.1%. Because of the small variance explanation of PPT results by education level (between 0.9% and 1.3%) and vascular risk factors (smoking, 0.4%; BMI, between 0.2% and 1.2%), only age- and sex-stratified normative data are presented (Table 3).
Normative Data for Men’s and Women’s Test Performance on Each Subtest by Age Group
Note. CI = confidence interval.
As can be seen from Table 3, PPT performance was quite stable until age 70 yr and then dropped on all subtests. Test results were lowest in the oldest age category, in which the difference from all younger age groups reached statistical significance for all subtests (p < .001). On all subtests, men performed significantly worse than women (p < .001). Participants with a history of diabetes also achieved significantly poorer results on all subtests (median of inserted pins = 14, 13, 11, and 26 for right hand, left hand, both hands, and assembly, respectively) than participants with no history of diabetes (median of inserted pins = 12, 12, 10, and 22 for right hand, left hand, both hands, and assembly, respectively; p < .001). Detailed statistics are given in Table 4.
Differences in the Scores on the Four PPT Subtests Between Four Age Groups, Men and Women, and Patients With and Without Diabetes
Note. PPT = Purdue Pegboard Test.
Kruskal–Wallace test.
Mann–Whitney U test.
Discussion
The main finding of our study is that older participants, men, and patients with diabetes achieved significantly lower scores on all PPT subtests than younger participants, women, and people without diabetes. PPT test performance declined significantly at age 70 yr and older compared with younger groups. The age-related decline in PPT results was seen in men and women, but women outperformed men in each age category. Despite the fact that diabetes was a significant determinant, it explained only a very little of the variance of PPT results in our study population.
Our findings are consistent with those of previous studies on manual dexterity (Agnew et al., 1988; Lawrence et al., 2014; Rule et al., 2021), showing an age-related deterioration not only in complex bimanual coordination tasks requiring effortful processing (Bangert et al., 2010; Fling et al., 2011), but also in the conduct of simple tasks, such as unimanual insertion of pins. We can only speculate on the underlying causes of declining manual dexterity with age. Both peripheral and central nervous system changes occurring with age have been implicated. Among peripheral changes, a decline in tactile sensitivity (Tremblay et al., 2003) may play a role. Other peripheral causes include the age- related reduction in muscle mass and in the number of motor units in the hands (Bangert et al., 2010). It is unlikely, however, that peripheral changes alone accounted for the decline in PPT performance in the older age groups in our study. A study by Dayanidhi and Valero-Cuevas (2014) supports this assumption, because it did not show associations between control of fingertip force and performance in peg-inserting tasks. Presumably, an important nonperipheral factor is the age-related decline in cognitive functioning. Cognitive abilities that are of particular importance for manual dexterity and are known to deteriorate with aging (Craik & Bialystok, 2006) are selective and divided attention, working memory, and executive functions (Drag & Bieliauskas, 2010; Reuter-Lorenz & Park, 2010; Zanto & Gazzaley, 2019). Among PPT subtests, we observed the strongest slope in decline for assembly. It is noteworthy that the assembly task differs from all other subtests because switching of actions is mandatory. These abilities are considered to belong to the executive function domain, and previous studies have reported that inhibition and switching are the first executive function deficits to occur during the course of aging (Carmeli et al., 2003; Jurado & Rosselli, 2007).
Although vascular brain changes are a major determinant of age-related decline in executive functions, it is unlikely that such brain abnormalities played a predominant role in the worsening of manual dexterity with aging in our population. Indeed, vascular risk factors explained only a very small proportion of the variability of PPT results among study participants. With regard to the poor PPT results among patients with diabetes, we assume that diabetic peripheral neuropathy, which affects the sensory function of the hands, could have negatively affected hand dexterity (Lima et al., 2017).
Our finding that women performed better than men on all four subtests of the PPT is in line with the findings of some previous studies (Desrosiers et al., 1995; Lezak et al., 2012), but not all (Rule et al., 2021; Vasylenko et al., 2018). The simplest explanation for the observed sex differences in manual dexterity is hand size. Manipulation of small pegs, such as those used in the PPT, is easier with women’s hands, which, on average, are smaller than men’s hands. Indeed, Peters et al. (1990) showed that the sex differences in dexterity performance disappeared if statistical analyses corrected for index finger and thumb thickness. Some authors explained the superior performance of women on dexterity tests as being due at least in part to women’s greater practice in household activities, many of which involve fine manipulation of objects (Merritt & Fisher, 2003). Others suggested that differences in brain aging between the sexes, such as a protective effect of estrogen on neurons and glial cells (Garcia-Segura et al., 2001), are responsible. If this were true, the difference in test performance between men and women in our study should have increased with advancing age. However, it remained almost identical over all age groups.
Strengths and Limitations
Our study has several strengths and limitations. Strengths are the large sample of community-dwelling participants without signs and symptoms of neurological disease. This extends previous work, which has often examined convenience samples or samples of individuals with minor clinical impairment. Limitations are the lack of study participants in the age groups younger than age 40 yr and older than age 80 yr and the lack of elaborate assessment of visual perception.
Future Research
Our data may also serve as a basis for longitudinal studies evaluating the role of motor function as a predictor of future disease. Among these diseases are dementing illnesses and movement disorders (Darweesh et al., 2017).
Implications for Occupational Therapy Practice
Dexterity tests such as the PPT measure the accuracy of hand and finger movements under controlled conditions. Test norms are helpful for occupational therapists in developing rehabilitation plans and measuring the effectiveness of such programs.
Conclusions
The results of current investigation are important because we provide normative data for the PPT stratified by age and sex in the 40- to 80-yr age range. In a clinical and research setting, these data can be used to classify the manual dexterity of participants according to the percentile distribution of a normal population. Such information is crucial when comparing groups with various health conditions.
Footnotes
Acknowledgments
We acknowledge the Medical University of Graz and the Sustainable Health Research Doctoral School.