Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Most statistical analyses are largely concerned with the appropriateness of the regression model used. However, for the data at hand, an important and often neglected problem concerns the issue of misclassification error, which leads to bias and loss of efficiency in estimation. It is possible to correct for this bias if the misclassification probabilities can be estimated unbiasedly from, for instance, the validation data. In this thesis, frequentist and Bayesian methodologies have been developed to correct for misclassification errors in discrete data, accounting for the uncertainty by which the correction terms are estimated. The main advantage of the Bayesian approach over the frequentist method for misclassification problems is that it allows for conceptually straightforward estimation, utilizing all the available information and providing an elegant way for taking the variability of the correction terms into account. The approaches that have been developed assume that along with the possibly misclassified main data measurements, error-free validation data measurements are available. Parameter estimation is done simultaneously combining information from both measurement processes. Most statistical analyses are largely concerned with the appropriateness of the regression model used. However, for the data at hand, an important and often neglected problem concerns the issue of misclassification error, which leads to bias and loss of efficiency in estimation. It is possible to correct for this bias if the misclassification probabilities can be estimated unbiasedly from, for instance, the validation data. In this thesis, frequentist and Bayesian methodologies have been developed to correct for misclassification errors in discrete data. The main advantage of the Bayesian approach over the frequentist method for misclassification problems is that it allows for conceptually straightforward estimation, utilizing all the available information and providing an elegant way for taking the variability of the correction terms into account. The analysis of caries research, particularly with application to the Signal Tandmobiel® study, revealed that the correction for misclassification error matters and the effect of correction varies depending on the type of the discrete response (binary, ordinal or count), and the associated model assumptions. It has been established that the East-West gradient in caries experience was not influenced by the potential misclassification of the dental examiners.

Choose an application

• Reference Manager
• EndNote
• RefWorks (Direct export to RefWorks)

Stroke is a major health burden throughout Europe and consumes a large amount of health care resources.1 Great differences exist in stroke management and outcome in Europe. Optimal models for delivery of stroke care, resulting in optimal outcome at reasonable cost are of great importance.1 Our understanding of the components of inpatient stroke rehabilitation that are critical for patients’ outcome is still limited. Comparing practice and outcome across European countries may give clues that can help to develop new hypotheses and intervention strategies. Therefore, longitudinal long-term follow-up studies are needed to monitor the progress of individual patients, to assess the performance of hospital and community services and to evaluate intervention.2 Optimally, such research should take place in a multi-centre study in a collaborative framework. The overall aim of this doctoral thesis was to identify components of inpatient stroke rehabilitation that have an impact on patients’ recovery. Therefore three studies were set up in four European rehabilitation centres (BE, UK, CH, DE). The findings of these studies are summarized below. In study 1 , the use of time of stroke patients while in the rehabilitation centre was documented. In each centre, 60 patients were monitored during 30 time sampling sessions: 10 morning (7.00 am-12.00 noon), 10 afternoon (12.00 noon-5.00 pm) and 10 evening sessions (5.00 pm-10.00 pm). Therapy time was minimal in the evening session. Therefore, the data of the evening sessions were not included in statistical analysis. The data of morning and afternoon sessions were analyzed with a generalized estimating equation model, controlling for serial dependency of the data and for confounders as age, initial motor and functional deficit. The main significant finding was that patients in the UK centre spent significantly less time in therapy compared to the other centres. The average absolute time in therapy was 1 hour in the UK centre, 2 hours in the Belgian centre, 2 hours 20 minutes in the German centre and 2 hours and 46 minutes in the Swiss centre. Low therapy time in the UK centre was in contrast to the time available for patients from all professional groups. A possible explanation may be the different division of tasks. Participant observations indicated that professionals in the UK centre spent more time in legally required administration, leaving less time for patient care. In all four centres, physiotherapy and occupational therapy together accounted for more than half the total therapy time. Compared with the other centres, patients in the UK centre spent less time in occupational therapy, but received more nursing care. This was probably the results of the high nurse staff levels. Sports-related activities and autonomous exercising were rarely observed in any centre, suggesting a potential for self-directed remedial therapy. Overall, this study revealed that in the participating centres, stroke patients spent a large amount of the day in their rooms, inactive, and without any interaction. Sitting, lying and sleeping accounted for a third to half of the day. The evidence that more intensive rehabilitation improves outcome after stroke3 was not reflected in the rehabilitation practice observed. Still, great differences occurred across the centres with patients in the UK and Belgian centres spending less time in therapy compared to patients in the Swiss and German centres. The latter centres had a more structured rehabilitation program. This may have resulted in more therapy time and a more challenging environment for the patients, physically and mentally. Therapy time devoted to stroke patients in a rehabilitation setting seems to be more dependent on the management style than on the number of staff available. In study 2 , the content of physiotherapy and occupational therapy for stroke patients was compared across the centres. First a scoring list was developed to define the content of individual physiotherapy and occupational therapy sessions for stroke patients. The therapeutic categories of the list were based on previous lists,4,5 neurological textbooks of stroke rehabilitation6 and existing videotapes of physiotherapy and occupational therapy sessions with stroke patients, made in different European rehabilitation centres. The list was finalized considering the suggestions of five physiotherapists and five occupational therapists, who had more than two years of experience in the field of neurological rehabilitation. The final scoring list consisted of 12 categories and 46 subcategories. An inter-rater reliability study was carried out with the four researchers of the different centres. Comparing the frequency of occurrence of the categories resulted in intra-class correlation coefficients, indicating high reliability for eight categories, good reliability for one category, and fair for two categories. One category was not observed. The developed scoring list was a helpful and reliable tool to unravel and compare the content of individual physiotherapy and occupational therapy sessions for stroke patients in inpatient rehabilitation settings in various European countries. The results encourage further use of the list in future research and practice aiming to improve evidence-based stroke rehabilitation. In a second phase, the list was used to compare the content of physiotherapy and occupational therapy sessions between the four rehabilitation centres. Every researcher scored 15 physiotherapy and 15 occupational therapy sessions, recorded in their own centre. Patients with different impairments might receive different treatments. Therefore therapy sessions were recorded from stroke patients fitting predetermined clinical criteria. This was done to cover the full spectrum of potential disabilities and to ensure an equivalent patient group in each centre. Additionally the data was pooled over the centres to compare the content of both therapeutic disciplines. Data were analyzed using a generalized estimating equation model controlling for serial dependency of the data and for confounders as age and duration of the treatment session. Comparison of physiotherapy and occupational therapy between centres revealed significant differences for only two of the twelve categories. Ambulatory exercises occurred more often in the physiotherapy sessions in the Belgian and UK centres and relearning selective movements occurred less in the physiotherapy and occupational therapy sessions in the UK centre. Comparison of the two therapeutic disciplines on the pooled data of the four centres, revealed that ambulatory exercises, transfers, exercises & balance in standing and lying occurred significantly more often in the physiotherapy sessions. ADL, domestic and leisure activities and sensory, perceptual training & cognition occurred significantly more often in the occupational therapy sessions. This study revealed that the content of each therapeutic discipline was consistent between the four centres. Physiotherapy and occupational therapy proved to be distinct professions with clear demarcation of roles. In study 3, motor and functional recovery patterns were compared across centres. In the four centres, 532 stroke patients were recruited. On admission to the centre and at two, four and six months after stroke the Barthel Index (BI)7 and Rivermead Motor Assessment-Gross Function (RMA-GF)8 were assessed. At two, four and six months, also the Nottingham Extended Activities of Daily Living (NEADL)9 was assessed. The statistical comparison of the recovery patterns across centres over time required an adjustment for case-mix and a mechanism for handling missing data (intention-to-treat analysis). Two other issues that complicated the comparison were the skewed distribution of the outcome variables and the earlier baseline measurement in the UK centre compared to the other centres. Therefore random effects ordinal logistic models were used for the analysis. The results showed that patients in the UK centre were significantly more likely to stay in lower RMA-GF classes compared to patients in the German centre. In the Swiss centre, patients were significantly less likely to stay in lower NEADL classes compared to patients in the UK centre. These findings should be interpreted in view of the previous studies. In chapter 1, we found that overall therapy time in the UK centre was significantly less compared to the other three centers.10 Also time in occupational therapy was significantly less in the UK, compared to the Swiss centre. In all centres, physiotherapy and occupational therapy comprised more than 50% of therapeutic time. In the UK centre, 35% of therapy time consisted of nursing care, which was more than in the other centres. In chapter 2, we reported that the content of physiotherapy and occupational therapy was consistent over the centres. The higher input of therapy in the Swiss and German centres was not related to higher staffing levels, but to a different time allocation of therapists and a strictly timed rehabilitation program for patients and therapists. This formal management led to a higher input of therapy, which in turn resulted in better motor and functional recovery for the patients. In contrast to the results for RMA-GF and NEADL, patients in the UK centre were significantly less likely to stay in lower BI classes compared to those in the German centre. This might be the result of the ceiling effect of the BI, the higher input of nursing care in the UK centre, the emphasis on self care to enable early discharge and the fact that middle band patients can expect more functional gain.11 The recovery patterns of the Belgian patients did not differ significantly from patients in any other centre. In summary, motor and functional recovery in the Swiss and German centres was better compared to the UK centre, with exception of self care recovery in the UK centre. In the German and Swiss centres, patients received noticeably more therapy per day. This higher therapy input was not a consequence of higher staffing levels, but of a more efficient organization of rehabilitation services. This study indicates a potential for further improvement of the services in the UK and Belgian centres without additional cost. In study 4 , the prevalence and predictors of post-stroke affective disorders were documented. Post-stroke depression and anxiety were assessed with the Hospital Anxiety and Depression scale (HADS)12 at two, four and six months after stroke. Based on the original publication, a score >7 on HADS-depression or anxiety subscale (range 0-21) was considered to reflect the syndrome of depression or anxiety, respectively.12 Prevalence and severity were compared across centres using Chi² and Kruskal-Wallis tests, respectively. Predictors and time course of severity of depression and anxiety were examined using linear mixed models on the pooled data. Of the 532 patients enrolled in the study, the HADS was not completed at any time for 27 patients. Consequently, 505 patients were included in the analysis. Overall the prevalence of depression at the several evaluation points varied between 21% and 39%. The overall prevalence of anxiety ranged between 15% and 30%. There was no significant difference in the prevalence or severity of both affective disorders between centres. Therefore we pooled the data from the four centres to examine if the high proportion of depressed and anxious patients at each time point comprised the same individuals. Results showed that patients reporting an affective disorder at six months comprised only half those with an onset before two months. The other half had a later onset. Linear mixed models analyses showed that stroke severity, functional disability, motor impairment and baseline sensory deficit were univariate predictors of severity of both depression and anxiety. Additionally baseline cognitive disorder, dysarthria, pre-stroke Barthel index and age were associated with the severity of depression. In the multivariate models only the initial Barthel Index was retained. After correction for predictive factors, levels of depression were stable over time, while anxiety levels decreased slightly. In conclusion, this study showed that the prevalence and severity of affective disorders after stroke was similar in the four European centres. Monitoring for affective disorders is crucial as many patients risk becoming depressed or anxious in the first 6 months after stroke. The multivariate models suggest a relationship between emotional distress and functional disability after stroke. References 1. Markus H. Variations in care and outcome in the first year after stroke: a Western and Central European perspective. J Neurol Neurosurg Psychiatry. 2004;75:1660-1661. 2. Hewer RL. Outcome measures in stroke. A British view. Stroke. 1990;21(9 Suppl):II52-II55. 3. Kwakkel G, van Peppen R, Wagenaar RC, Wood Dauphinee S, Richards C, Ashburn A, Miller K, Lincoln N, Partrdige C, Wellwood I, Langhorne P. Effects of augmented exercise therapy time after stroke: a meta-analysis. Stroke. 2004; 35:2529-2539. 4. Ballinger C, Ashburn A, Low J, Roderick P. Unpacking the black box of therapy – A pilot study to describe occupational therapy and physiotherapy interventions for people with stroke. Clin Rehabil. 1999;13:301-309. Gladman JHF, Juby LC, Clarke PA, Lincoln NB. Survey of a domiciliary stroke rehabilitation service. Clin Rehabil. 1995;9:245-249. 5. Davies PM. Steps to follow. The comprehensive treatment of patients with hemiplegia , Second edition, Springer, 2000. 6. Mahoney FI, Barthel DW. Functional evaluation: the Barthel Index. Md State Med J. 1965;14:61-65. 7. Lincoln N, Leadbitter D. Assessment of motor function in stroke patients. Physiotherapy. 1979; 65:48-51. 8. Nouri FM, Lincoln NB. An extended activity of daily living scale for stroke patients. Clin Rehabil. 1978;1:301-305. 9. De Wit L, Putman K, Dejaeger E, Baert I, Berman B, Bogaerts K, Brinkmann N, Connell L, Feys H, Jenni W, Kaske C, Lesaffre E, Leys M, Lincoln N, Louckx F, Schuback B, Schupp W, Smith B, De Weerdt W. Use of time by stroke patients: a comparison of four European rehabilitation centres. Stroke. 2005 . 36:1977-83. 10. Alexander MP. Stroke rehabilitation outcome. A potential use of predictive variables to establish levels of care. Stroke 1999;25:128-34. 11. Zigmond AS, Snaith RP. The hospital anxiety and depression scale. Acta Psychiatr Scan. 1983;67:361-370. 12. Horn SD, DeJong G, Ryser DK, Veazie PJ, Teraoka J. Another look at observational studies in rehabilitation research: going beyond the holy grail of the randomized controlled trial. Arch Phys Med Rehabil. 2005;86:S8-S15. Vergelijkende studie van de revalidatie van patiënten met een beroerte in vier Europese landen Het cerebrovasculaire accident (CVA), of beroerte is een belangrijk gezondheidsprobleem in Europa. De multidisciplinaire behandeling van CVA-patiënten in gespecialiseerde ziekenhuisafdelingen resulteert in een daling van het aantal sterfgevallen en verminderde fysieke beperkingen van de patiënt. Er blijken binnen Europa echter grote verschillen te bestaan in de mate van herstel van patiënten met een beroerte. De precieze aspecten van de multidisciplinaire aanpak, die van cruciaal belang zijn voor het herstel zijn onvoldoende gekend. Studies waarbij zowel het herstel van patiënten als de aangeboden dienstverlening gedocumenteerd worden, zijn nodig om onze kennis te verdiepen. Het doel van dit doctoraatsproject was de revalidatieaanpak en de herstelpatronen van CVA-patiënten te vergelijken tussen vier revalidatiecentra die typisch zijn voor de vier Europese landen. De deelnemende centra waren: Universitair Ziekenhuis, Pellenberg (België), City Hospital en Queen’s Medical Centre, Nottingham (Groot-Brittanië), RehaClinic, Zurzach (Zwitserland) en Fachklinik, Herzogenaurach (Duitsland). De resultaten van deze studie tonen aan dat het motorische en functionele herstel van CVA-patiënten minder goed verloopt in het Britse revalidatiecentrum in vergelijking met het Duitse en Zwitserse centrum. De kinesitherapie en ergotherapie die in de vier centra wordt aangeboden is in hoge mate vergelijkbaar. In het Britse centrum ontvingen patiënten echter beduidend minder therapie. Eigenaardig genoeg waren deze verschillen in hoeveelheid aangeboden therapie niet gerelateerd aan verschillen in personeelsbezetting. Het bleek echter dat men in het Duitse en Zwitserse centrum een meer formele en strikte managementstijl hanteerde. Dit resulteerde in een efficiëntere tijdsbesteding van de therapeuten.