Parkinson's disease (PD) is a progressive neurodegenerative disorder, characterised by bradykinesia, tremor, disturbances of postural reflexes and of the autonomic nervous system, and which is most prevalent in old age 123.

Diagnosis and management of PD is based on clinical assessment of symptoms and signs using disease specific rating scales 45. Evaluation of disease severity, effect of treatment procedures and interventions are all influenced by the patient's perception of their disease and assessment instruments have to take this aspect into account 456. Such instruments measuring different factors affecting patient's perception of well-being in relation to disease and health are referred to as health related quality of life (HRQOL) questionnaires. Measures of general HRQOL, like the SF-36 78 are used widely but are less suited to assessment and management of a specific health condition compared to their disease specific counterparts, like the 39-question Parkinson's Disease Questionnaire (PDQ-39) 9.

The PDQ-39 has been translated into many languages all over the world and extensively used for PD research by many authors 10111213. Translation and validation of an instrument is a prerequisite for its use in a new cultural context and only after validation is it possible to compare the research data in a cross-cultural context 1415.

The authors hypothesized the Estonian version PDQ-39 to be an acceptable, valid and reliable measure for use in PD patients in Estonia and the aim of current study was to test these hypotheses.

Psychometric analysis was conducted in two groups of patients: 15 patients in the pilot study for processing the qualitative assessment of the questionnaire and 104 patients in the main study group, for which 81 self-reported questionnaires were completed and used in the statistical analysis, a response rate of 79%. The PD patients who entered the final analysis group were categorised by severity of disease as follows: 48 patients (59%) in mild stage (HY stage I–II), 29 patients (36%) in moderate stage (HY III) and 4 patients (5%) in severe stage (HY IV–V). Non-responders belonged to mild (16 patients) and moderate (5 patients) disease groups while two disease severity assessment questionnaires were unusable. Descriptive characteristics of the responders and non-responders are shown in Table 1.

The most frequently reported concomitant diseases were high blood pressure (24%), coronary heart disease (24%), and joint disease (14%). There were two cases of stroke and one of malignancy among the patients in the severe disease group. The majority (60%) of patients lived in an urban area. The most common medication was levodopa and all patients with severe PD suffered from adverse effects of their medications.

Validity interpretation was carried out using the ability of the PDQ-39 to distinguish high and low symptom burden and correlation of questionnaire responses to clinically assessed severity of PD. The patients in mild stages of PD had an average summary index score 35.8, for patients in a moderate stage of disease it was 46.1 and for patients in an advanced stage of the disease it was 59.1 (Figure 1). There was a statistically significant difference between the domain scores in patients with mild versus moderate PD for the domains of Mobility, ADL, and Communication. The difference in scores for Stigma, Social support and Cognition was not significant.

Content validity was assessed through the qualitative content analysis during the pilot study. All the questions and domains were essential to PD patients for measurement of quality of life.

Internal consistency for all domains was good as the Cronbach alpha measure was above 0.7 in all cases (ranging from 0.81 to 0.86) indicating good reliability of the questionnaire (see details in Table 2). Item-test correlations for all questions and domains ranged from 0.56 to 0.83 again confirming the consistency of the questionnaire (only domain figures are given in Table 2 for the sake of brevity). The item-item correlations of mean domain scores were moderate ranging from 0.67 between Mobility and Activities of Daily Living to 0.2 between Communication and Bodily Discomfort (Table 3) as expected.

Reproducibility of the results and test-retest reliability of the questionnaire were assessed by correlation between results of test and retest questionnaires from a selection of patients. These results ranged from 0.72 to 0.92 indicating good reproducibility. The retest number of responders was 78.

The whole range of possible scores (0–100) of the PDQ-39 domains were covered. The mean and median were very similar in the domains of Mobility, Activity of Daily Living (ADL), Emotional Well-Being and Stigma, with a difference of less than 10%. For Cognition, the difference was 16%. The difference between the mean and median for the summary index was approximately 1% (see Table 4 for details).

The floor effect ranged from 1.2% to 2.5% in the domains of Mobility, ADL, Emotional Well-being, Cognition and Bodily Discomfort, while for the domains of Stigma, Social Support and Communication it ranged from 16% to 24%, which was more than hypothesized. It is also noteworthy that among patients with moderate or severe PD patients the response distribution was more central and the only floor effect was recorded for the Social Support domain (21% compared to the 25% in the mild PD group). The lowest ceiling effects (1.2%) were observed in Mobility, ADL, Emotional Well-Being, Social Support and Communication domains. In other domains the ceiling effect was between 2.4% to 3.7%. The criterion of ceiling effects below 15% was achieved in all domains. Both ceiling and floor effects of the Summary index (SI) were at the 1.2% level. The analyses showed that the distribution of the results in the domains and summary index was within the limits of the preset criteria in most cases.

This paper presented the evidence on acceptability, validity and reliability of the Estonian version of the PDQ-39 which were good as originally hypothesised. The descriptive statistics for the acceptability assessment of the questionnaire showed a good distribution of the response choices on the scale level, ceiling and floor effects were low and within hypothesized limits with only few exceptions. Content and face validity tests unambiguously showed that the covered topics and items in the translated PDQ-39 questionnaire are relevant to the Estonian PD patients. Validity assessment of the Estonian version of the PDQ-39 was guided by experience from previous similar studies 62122232425 and as expected the localised version of PDQ-39 was able to differentiate between disease groups with the lowest and highest PD burden. Most of the previous validation studies in other cultural settings 910111213141516 have used a variety of rating scales and classifications like Hoehn and Yahr stages, UPDRS scores and main symptoms for assessment of construct validity. All these studies found statistically significant differences between disease severity and PDQ-39 domains as an example for the domains of Mobility and Activities of Daily Living (mean scores ranging from 22.34 (HY I) to 75.25 (HY IV–V) 6222425 which ranged from 36.5 (HY I–II) to 90.5 (HY IV–V) in our study. The differences of the domains of Stigma, Social support, Emotional Well-Being and Bodily discomfort were not so clearly related to the severity of PD stages. The summary index scores were the lowest in the patient group with mild PD and highest in the severe disease group as expected to once more indicating close psychometric values to other language versions of PDQ-39 111222. The somewhat worse domain scores especially in the case of Mobility and Activities of Daily living could be related to the still developing social care system in Estonia but verification of this assumption would recuire a direct comparison of country situations and respective social care systems.

Regarding the reliability and other internal consistency measures the strongest correlations between domain and total questionnaire score have been found for the domains of mobility and Activities of Daily Living 910111213141516. These two domains showed strong item-test correlations in our study as well. The reliability figures of Estonian version of PDQ-39 surpassed the results from cross-cultural studies from USA, Canada, Spain, Italy and Japan 10 as witnessed by Cronbach alpha above 0.80 26 for all domains compared to a wider range (from 0.13 to 0.96) in those studies. The lowest Cronbach alpha coefficients in those studies were found in Japan and seemed to be connected to that particular cultural background 10. Test-retest analysis showed no differences of score distributions between the two assessments as no changes were recorded in previous reports of the original version of PDQ-39 152224. The localisation of PDQ-39 to Estonian context was successful since all preset criteria were met and the results of this study were comparable to the results of previous validation studies in other cultural settings in UK, USA, Canada, Spain and Italy.

One of the possible drawbacks of the current study is associated with the relatively small sample size. However, during the study planning a relative size of the study group was set at 5% of the total PD population in Estonia and that target was reached with the current sample. Also, the mean age of the patients in the sample corresponded well to the mean age of the overall PD population in Estonia 27. Especially good representativeness of the Estonian PD population was achieved for the patient groups with mild and moderate disease. However, the patient group with severe PD had better response rate than the group with mild PD as 4 out of 5 patients did return a filled questionnaire. The PDQ-39 is a self-administered written questionnaire and the study inclusion criteria therefore were set to include only patients having the ability to perform that particular task. The small number of patients with severe PD available for validation of written PDQ-39 highlighted the need for a HRQL measure in Estonia that could be administered in an interview form to the severe patients of PD.

The analysis of sample variation and statistical significance testing showed that albeit the possible drawbacks related to the sample size and composition the Estonian language version of PDQ-39 performed well in the psychometric testing. Therefore, the use of PDQ-39 is highly warranted for disease assessment and treatment management among the PD patients able to administer the questionnaire as the localised version of the PDQ-39 is a valid and reliable tool in this setting.

The Estonian version of the PDQ-39 demonstrated similar validity and reliability to the original English language version.

It is an acceptable, valid and reliable instrument to measure quality of life of PD patients in future studies in Estonia.

PD – Parkinson's disease; PDQ-39 – Parkinson 's disease Questionnaire; HRQOL – health related quality of life; ADL – Activities of Daily Living; MMSE – Mini Mental State Examination; HY – Hoehn and Yahr staging system; UPDRS – Unified Parkinson's Disease Rating Scale version 3.1.

The author(s) declare that they have no competing interests.

ÜK collected data and drafted the whole manuscript. TA was involved in conception and design the study. PT contributed in interpretation of data and in selection of patients, TL performed statistical analysis and contributed in interpretation of data. All authors read and approved the final manuscript.