No full-text available
To read the full-text of this research,
you can request a copy directly from the authors.
Regression zur Mitte
Abstract
Abstract
Regression to the mean
The evaluation of rehabilitation programs may be distorted by regression to the mean: In a group of patients with extreme measurement values, these values tend to be less extreme on a following point in time due to merely random components and regardless of a ‘true’ treatment effect. If this effect is not taken into account, the effectiveness of rehabilitation programs may be estimated wrongly. In this paper regression to the mean is explained comprehensively and common misunderstandings are clarified. It is shown, which conditions are crucial for regression to the mean to arise and which factors determine its strength. Furthermore it is shown how regression to the mean should be controlled in repeated measurement designs.
... Finally, the better performance of the improvers group in the afternoon compared to the morning could be also attributed to a possible selection bias in the sense of a "regression to the mean," which occurs when selection has taken place in successive series of measurements that are linearly related and not perfectly correlated, resulting in more or less extreme baseline values and may ultimately lead to incorrect conclusions (Zwingmann & Wirtz, 2005). In this regard, the "improvers'" better performance in the afternoon could be interpreted as a mechanism of normal variability. ...
Objective: Fatigue in multiple sclerosis (MS) is common, burdensome, and usually assessed by self-report measures. This retrospective data analysis of the twice-daily Alertness test (Test battery of Attentional Performance) examined the extent to which this assessment procedure is associated with MS-related fatigue.
Method: Two-hundred and thirteen German inpatients (136 women) aged 18–69 years with predominantly relapsing MS (72.8%) were included. Based on reaction time (RT) differences between morning tonic alertness (8:30–11:00 a.m.) and
afternoon tonic alertness (3:00–4:30 p.m.), patients were divided into an “improver,” “maintainer,” or “decliner” group. Multinomial logistic regression (MLR) was calculated to predict the likelihood of belonging to one of these performance groups, taking into account cognitive fatigue (Fatigue Scale of Motor and Cognition, FSMCcog), disease severity (Expanded Disability Status Scale, EDSS), depression (Center for Epidemiologic Studies Depression Scale, CES-D), gender, and tonic alertness (a.m.).
Results: The final MLR model (R2 = .30) included tonic alertness (a.m.) (<.001), FSMCcog (.008), EDSS (.038), CES-D (.161), and gender (.057). Using this model, correct assignment to alertness performance groups was 56.8%. Tonic alertness
(p.m.) demonstrated the greatest potential for differentiation among the three performance groups (<.001).
Conclusions: These results show a relationship between subjective fatigue and tonic alertness. However, other variables also contribute to this association, suggesting that the RT differences between twice-daily measures of tonic alertness is not related to increased subjective fatigue in a substantial number of pwMS, which diminishes the diagnostic value. Further studies including relevant variables such as sleepiness are urgently needed.
... The tendency of declining agreement with wisdom sentences among students after the lecture was surprising in the results of our study. A possible explanation may be regression to the mean, which cannot be excluded in repeated-measures surveys, in the case of initially above-average values (Zwingmann & Wirtz, 2005): Students expressed higher agreement than the wisdom lecture group in the first wisdom rating. It is also conceivable that some fatigue, reluctance, or impatience occurred after the 90-minute diagnostic lecture (psychology stu-dents are routinely asked to complete questionnaires) and influenced responses to the second wisdom questionnaire. ...
Wisdom is the "ability to solve unsolvable problems" in life. Empirical studies have shown that wisdom can be strengthened by training. The present study investigates whether general population persons change their agreement to wisdom principles after a short informational lecture on "wisdom and wisdom skills." The development of wisdom attitudes in this wisdom lecture group (WG, n=39) is compared to wisdom agreement of a control group which did not receive wisdom information (CG, n=44, students of a lecture without wisdom-related content). Agreement with wisdom principles is assessed before and after the wisdom lecture (WG) or before and after the non-wisdom lecture (CG) using the 12-WD wisdom attitudes scale.
After the wisdom information lecture, participants report stronger agreement with wisdom principles than before (Mpre=7.43 - Mpost=7.56, scale 0-10). It is possible that an increase in knowledge strengthens the tendency to agree. In contrast, students who heard a non-wisdom-related lecture had a decrease in agreement (Mpre=8.00 - Mpost=6.19). Agreement to wisdom principles was independent of age, gender, qualification degree. The study gives reason to consider changeability of wisdom attitudes through information. Randomized controlled experimental studies are needed for further evaluation of wisdom interventions for the general population.
... One could argue that the change in teacher self-efficacy resulted due to regression to the mean and, therefore, is estimated wrongly [cf. (62)]. In that case, it would have been more likely to observe that high teacher-self efficacy decreased preand post-intervention, instead of remaining constant. ...
Introduction: Teaching is considered a mentally challenging occupation. Teacher self-efficacy is a personal resource which buffers the experience of stress and may be important in maintaining mental health. The preventive intervention “Manual-Based Psychological Group Program for Teachers” (MBPGPT) was applied and evaluated state-wide to improve the mental health of teachers. This study aims to investigate the intricate relation between teacher self-efficacy and mental health and their changes in the course of the intervention.
Method: Using a single-group pre-/post-design, the relation between teacher self-efficacy and mental health was investigated in 742 teachers. Pre- and post-changes in teacher self-efficacy and their interaction with mental health were examined in a subsample of 171 teachers, who met the conservative inclusion criteria. In ancillary analyses, correlations with underlying changes in work-related behavior and experience patterns were analyzed to better understand the intricate link between teacher self-efficacy and mental health.
Results: Teacher self-efficacy showed a significant, moderate correlation with mental health. Self-efficacy was moderately higher after the intervention than before the intervention, but independent of changes in mental health. Teacher self-efficacy was related to work-related psychological resistance and positive emotions. An increase in teacher self-efficacy was accompanied by an improvement in life satisfaction and distancing ability. A decrease in teacher self-efficacy went hand in hand with reduced experience of social support.
Discussion: This study confirmed teacher self-efficacy as an important, reliable resource and its correlation with psychological resistance. The absence of a control group limits what causal conclusions can be drawn from the study. Nevertheless, self-efficacy seems to be a worthwhile goal of preventive interventions for teachers and should be promoted due to its wide-ranging implications. Suggestions for further studies and interventions are made.
... In an experimental set up providing data that may depend on many variables, chance can be involved and it is to be ruled out that extreme outcomes observed at one point in time are followed by more moderate ones at another by chance and not due to a real improvement. If regression to the mean were the case, the measures observed in our patients prior to mPT would not differ from measures previously reported in healthy populations and the standard deviations in all analyzed subpopulations would be larger at T2 compared to T1 (68). However, our here analyzed data set did not fulfill these criteria for regression to the mean and allowed the comparison of hTNF and lTNF patients under real life conditions in the light of published data on healthy cytokine production by PBMC. ...
Objective: In experimental settings, systemically elevated inflammation markers interfere with major depression treatment. In German healthcare, compulsory national health insurance covers treatment of a wide variety of depressive disorders, if it follows evidence-based medicine guidelines combining recommended therapies. To date, little is known about the relevance of immune system cytokine production with regard to real-world clinical care for patients with moderate depression.
Methods: Seventy three patients with moderate depression subjected to multimodal psychotherapeutic inpatient therapy (mPT) following a psychodynamic concept at a German university hospital were included. As a primary outcome, mPT success, evidenced by delta HADS “depression,” was analyzed according to tumor necrosis factor alpha (TNFα) production by peripheral blood mononuclear cells (PBMC) after phytohemagglutinin (PHA) challenge at baseline. Secondary outcomes addressed the inflammatory response and mental health comparing high and low TNFα-producers.
Results: First, higher PBMC TNFα production at baseline predicted a better mPT-outcome (R² 0.162, p = 0.014). Second, patients with high TNFα (hTNF) at baseline produced significantly more acute inflammatory cytokines [interleukin (IL)1β, IL6), TH1/TH2 cytokines [interferon gamma (IFNγ), IL4] as well as eotaxin and IL2 compared to low TNFα producers (lTNF) (Cohen's ds between −0.532 and −1.013). Demographic data, diagnosis subtype-distribution, medication, systemic inflammation markers [C-reactive protein (CRP), high mobility group box 1 (HMGB1), leptin], anxiety and depression (HADS) did not differ. From baseline to mPT-discharge, HADS “depression” decreased in both hTNF (11.31 to 5.47, p = 0.001, d = 1.184) and lTNF patients (11.50–7.92, p = 0.001, d = −0.765), while PBMC cytokine production decreased significantly in hTNF (Cohen's ds between −0.304 and −0.345) with a significant group by time interaction for TH1/TH2 ratio. At the end of therapy, comparison of TNF groups revealed significantly lower depression-scores in hTNF compared to lTNF patients (5.47 compared to 7.92, p = 0.035, d = 0.504).
Conclusions: Our study demonstrates successful treatment of depression in a clinical care setting using multimodal psychotherapy based on a psychodynamic concept following guideline recommendation. The greatest improvement in patient depression was linked to the highest production of TNFα by PBMCs at baseline. Our study contributes to the definition of patient subpopulations with differing cytokine responses that are related to succesful treatment of depression.
... It is remarkable that-even though the diverging information interventions did not work out as intended for domain-general beliefs (see Section 4.1)-this pattern also emerged for domain-general beliefs. Of course, the observed pattern of effects in the diverging information groups could be explained (at least partially) by regression to the mean (e.g., Zwingmann & Wirtz, 2005). Regression to the mean should, Table 6 Regression coefficients of the exploratory target model predicting epistemic change in the D-Index (measured on a topic-specific and domain-general level). ...
Changing epistemic beliefs (beliefs about knowledge and knowing) requires individuals to experience epistemic doubt (a specific type of cognitive dissonance). To evoke epistemic doubt, many studies rely on presenting diverging information (conflicting evidence). However, not much is known about how different types of diverging information and individual differences affect epistemic change. In a preregistered study (N = 509), we investigated how interventions based on resolvable/unresolvable diverging information influenced epistemic change compared to non-diverging information. Moreover, we examined the role of prior epistemic beliefs in this regard. Multiple-group latent change analyses showed that topic-specific epistemic beliefs prospered in the two diverging information groups but not for non-diverging information, while domain-general beliefs remained largely unchanged. Although epistemic change was—as expected—more pronounced for individuals with naive prior beliefs in diverging information groups, detrimental effects existed for advanced prior beliefs. These findings point to the important role of prior beliefs in epistemic change.
... Rehabilitanden, die extrem belastet sind, werden in Referenz zu der Population aller Rehabilitanden bei einer zweiten Messung weniger extrem belastet erscheinen. Dies führt in der Regel zu einer Über- schätzung von Interventionseffekten, wenn keine adäquate Kontrollgruppe vergleichend betrachtet wird [7]. Designeffekte -verzerrende Effekte, die durch die Studiendurchführung bedingt sind: ...
Zusammenfassung
Um Effekte von Rehabilitationsmaßnahmen prüfen zu können, muss das Studiendesign so gewählt werden, dass Veränderungen der Zielvariablen (z. B. Gesundheitszustand) möglichst eindeutig auf die Interventionsmaßnahme zurückgeführt werden können. Bei Beobachtungsstudien muss davon ausgegangen werden, dass konfundierende Merkmale den Einfluss der Maßnahme auf die Zielvariable systematisch überlagern und Effektschätzungen verzerren. Die Präferenz für eine Behandlungsalternative, die Behandlungsmotivation sowie das Ausmaß der initialen Beeinträchtigung der Rehabilitanden sind typische Konfundierungen, die insbesondere bei freiwilliger Behandlungsauswahl oder bei Teilnahme nach Behandlungsempfehlung berücksichtigt werden müssen (programmexterne Störfaktoren). Bei gut kontrollierten Interventionsstudien (insb. randomisiert kontrollierte Studien, RCT), bei denen die Inanspruchnahme der Behandlung systematisch kontrolliert erfolgt, können diese Fehlerquellen vermieden werden. Aber auch bei gut kontrollierten Interventionsstudien kann die Studiendurchführung (z. B. aufgrund der Stichprobenauswahl, des Einsatzes von Messinstrumenten, des Austauschs zwischen Vergleichsgruppen; Designeffekte, Treatmentkontaminationen) die eindeutige Interpretierbarkeit der Studienbefunde beeinträchtigen. In diesem Beitrag wird gezeigt, welche Konfundierungen und Störeinflüsse typischerweise die Validität der Schlussfolgerungen aus empirischen Zusammenhängen verringern können. Es wird verdeutlicht, wie potenziell verzerrende Konfundierungen und Effekte von Störfaktoren durch die Gestaltung der Untersuchung vermieden oder bestmöglich kontrolliert werden können.
... Die Stichprobe mit den deutlichsten Verschiebungen von t1 zu t2 in der Angst-Skala ist die der orthopädischen Rehabilitanden, bei welcher die Erstmessung am Beginn der (für die Teilnehmer nicht unbedingt belastenden, aber doch sehr ungewohnten) Reha-Maßnahme durchgeführt wurde. Die Absenkung hin zu t2 ist mit dem Ausgangswertgesetz gut vereinbar (Zwingmann & Wirtz, 2005). Ebenso ist plausibel, dass die Werte der Allgemeinbevölkerung durch Boden-Effekte im Mittel etwas weniger streuen als die der Patientengruppen. ...
Objectives: Temporal stability of clinical-psychological questionnaires is problematic from a conceptual point of view, and it is empirically insuffi-ciently examined. In this paper we analyze the temporal stability of anxiety and depression, based on large samples. Methods: In three samples with N > 300 each (cancer patients, orthopaedic patients, and general population) the stability of anxiety and depression was assessed with the Hospital Anxiety and Depression Scale (HADS). Results: The stability coefficients range from 0.60 to 0.75. They are highest in the sample of the orthopaedic patients with a time interval of only three weeks between the examinations. The temporal stability of anxiety and depression in females is as high as that of males. Individual change scores of anxiety and depression are correlated with about 0.50. Conclusions: The coefficients which are usually adopted in the measurement of change are not suited to describe temporal stability. Further research with sufficiently large samples is needed to unravel the conditions which determine the stability.
... However, another explanation for this effect might involve methodological issues, namely, that the questionnaire may not be as sensitive for low levels of fear as it is for high levels. In addition, the potential effect of statistical regression to the mean must be considered [42]. ...
Objective:
The aim of this study was to investigate fear of disease progression (FoP) during the year following diagnosis of breast cancer and its association with general self-efficacy (SE).
Methods:
In a prospective study, 118 breast cancer patients were recruited shortly after diagnosis disclosure (response rate: 54%) and at 1-year follow-up (follow-up rate: 90%). Participants completed self-report measures of general self-efficacy (General Self-Efficacy Scale) and fear of progression (short form of the Fear of Progression Questionnaire).
Results:
Cross-sectional regression analysis revealed that high FoP is significantly associated with low SE, even when controlling for demographic and medical characteristics (total R² = 0.17). Having children and a relatively short time since diagnosis also significantly predicted higher FoP. Longitudinal analyses showed that FoP decreased significantly over time (p = 0.001; d = 0.25), but a significant decrease was only observed for patients with high initial FoP (p < 0.001; d = 0.74) and not for those with low initial FoP (p = 0.688; d = 0.08). SE was not a significant predictor of FoP at follow-up when controlling for initial FoP and other patient characteristics (incremental R² = 0.001; p = 0.674; total R² = 0.47). Overall, only initial FoP significantly predicted FoP at follow-up (p < 0.001; β = 0.671).
Conclusion:
Findings that low SE is associated with high FoP can help to improve the treatment of dysfunctional fears in breast cancer patients. As FoP changes only slightly over time, treatment to enhance SE and reduce FoP should be initiated soon after disease disclosure.
... Medical rehabilitation programmes for example, often are evaluated for their ability to restore the patient's ability to work. Unaware of RTM effects a patient's recovery typically is interpreted as a treatment effect [2]. Other examples include the evaluation of asthma disease management programmes [3] or cholesterol screening [4]. ...
Regression to the mean (RTM) occurs in situations of repeated measurements when extreme values are followed by measurements in the same subjects that are closer to the mean of the basic population. In uncontrolled studies such changes are likely to be interpreted as a real treatment effect.
Several statistical approaches have been developed to analyse such situations, including the algorithm of Mee and Chua which assumes a known population mean mu. We extend this approach to a situation where mu is unknown and suggest to vary it systematically over a range of reasonable values. Using differential calculus we provide formulas to estimate the range of mu where treatment effects are likely to occur when RTM is present.
We successfully applied our method to three real world examples denoting situations when (a) no treatment effect can be confirmed regardless which mu is true, (b) when a treatment effect must be assumed independent from the true mu and (c) in the appraisal of results of uncontrolled studies.
Our method can be used to separate the wheat from the chaff in situations, when one has to interpret the results of uncontrolled studies. In meta-analysis, health-technology reports or systematic reviews this approach may be helpful to clarify the evidence given from uncontrolled observational studies.
... Dies unterstreicht die Notwendigkeit validierter und zuverlässiger, aber kompakter Messinstrumente, die in der Lage sind, die wichtigsten Aspekte der Krankheit und Krank− heitsbewältigung zu erfassen.Spontanremissionen, Regressionseffekte, TrendsEine weitere Einschränkung stellt die Möglichkeit dar, dass die in der Studiengruppe erzielten Verbesserungen nicht auf die Inter− ventionen, sondern (auch) auf andere Einflüsse zurückgehen. So ist bei den Krankheitstagen seit Jahren ein Rückgang zu beobach− ten (AU−Trend)[23]. Allerdings sollten beide Untersuchungs− gruppen in gleicher Weise von diesen Einflüssen betroffen sein: wie oben würden die Unterschiede zwischen den Gruppen durch solche Einflüsse nicht erklärt. Da sich die Ausgangswerte unserer Untersuchungsgruppen nicht unterscheiden, können auch (un− terschiedlich ausgeprägte) Regressionseffekte für die Unter− schiede zwischen den Gruppen weitgehend ausgeschlossen wer− den. ...
According to a recent review by Hüppe and Raspe effects of multidisciplinary treatment programs for patients with chronic low back pain in Germany seem to be rather weak and not to have persisting effects. Factors which could counteract possible benefits of treatment are, among others, psychic and job-related stresses and strains persisting after treatment. A multidisciplinary, in-patient treatment program for patients with chronic low back pain, therefore, was amended by multidisciplinary diagnosis and assignment and measures to support vocational solutions.
To evaluate the effects of the multidisciplinary program in comparison to a control group with the usual care, a prospective longitudinal study was conducted. 307 patients were assigned to the multidisciplinary in-patient treatment program, whereas 176 patients with comparable complaints had the standard rehabilitation program. Besides the full sample, we analyzed a subgroup of patients with chronic low back pain.
We found positive moderate and strong effects in the intervention group concerning function, pain, psychic strains as well as the number of sick days and return to work rates 10 months after discharge. Effects in the intervention group exceeded the effects achieved in the control group. Beside the full sample, we analyzed a subgroup of patients with chronic low back pain, who received an intense activating group treatment. Also in this subgroup we found moderate and strong effects of treatment superior to those in the control group for function, psychic strains and sick days.
We attribute these persisting and superior effects in the treatment group to an efficient treatment of occupational and psychic problems as well as to more homogeneous treatment groups attained by a multidisciplinary diagnosis and team-based assignment. They also show the significance of in-patient-treatment which is effective, when -- based on multidisciplinary diagnosis -- differential treatment groups can be formed.
... Dies unterstreicht die Notwendigkeit validierter und zuverlässiger, aber kompakter Messinstrumente, die in der Lage sind, die wichtigsten Aspekte der Krankheit und Krank− heitsbewältigung zu erfassen.Spontanremissionen, Regressionseffekte, TrendsEine weitere Einschränkung stellt die Möglichkeit dar, dass die in der Studiengruppe erzielten Verbesserungen nicht auf die Inter− ventionen, sondern (auch) auf andere Einflüsse zurückgehen. So ist bei den Krankheitstagen seit Jahren ein Rückgang zu beobach− ten (AU−Trend)[23]. Allerdings sollten beide Untersuchungs− gruppen in gleicher Weise von diesen Einflüssen betroffen sein: wie oben würden die Unterschiede zwischen den Gruppen durch solche Einflüsse nicht erklärt. Da sich die Ausgangswerte unserer Untersuchungsgruppen nicht unterscheiden, können auch (un− terschiedlich ausgeprägte) Regressionseffekte für die Unter− schiede zwischen den Gruppen weitgehend ausgeschlossen wer− den. ...
Chronic diseases of the musculoskeletal system rank first as causes of early retirement in Germany. Therefore orthopaedic rehabilitation has to identify patients with work-related problems and to promote return to work through differential treatment and vocational counselling. In the framework of the IopKo-Project such measures were developed and evaluated. These measures encompass: (1) an intensive and multiprofessional diagnostic pathway which allows early detection and treatment of mental disorders and job related problems; (2) homogeneous treatment groups based on multiprofessional diagnostics; (3) differential treatments, among these a multidisciplinary programme for patients with chronic low back pain or high risk of chronification (Rückenfit); (4) interactive training modules which mediate principles of performance and disability expertise, the legal bases of retirement pensioning, and measures to support occupational rehabilitation; and (5) a work hardening training programme.
To evaluate the effects of these measures in comparison to a control group with usual care, a prospective longitudinal study was conducted. A total of 307 patients were assigned to the multidisciplinary in-patient treatment programme, whereas 176 patients in the control group had a standard rehabilitation programme.
The results show positive moderate and strong effects in the intervention group concerning function, pain, psychic strains as well as the number of sick days and return to work rates 10 months after discharge. The effects in the intervention group exceeded the effects achieved in the control group. PATIENTS WITH HIGH RISK OF CHRONIFICATION: Beside the full sample, a subgroup of patients with chronic pain or high risk of chronification was analyzed, who had received a multidisciplinary functional restoration treatment. Also for this subgroup we found moderate and strong effects of treatment for function, psychic strains and sick days superior to those in the control group.
By this study we were able to show that orthopedic rehabilitation in a multimodal and multidisciplinary setting with a focus on activating and motivating therapy can have sustainable positive effects on pain, function and psychic well-being as well as on economic parameters. We interpret these persistent and superior effects in the treatment group (1) as a result of multiprofessional diagnosis and assignment which helps to subdivide the inhomogeneous group of patients with unspecific back pain into more homogeneous and thus more effective subgroups, (2) as a result of increased motivation by closed treatment groups, (3) as a result of intense and multilevel counselling of work related problems, (4) as a result of work hardening modules, and (5) as a result of direct and efficient treatment of psychic strains. The results also demonstrate the significance of inpatient rehabilitation, which will be efficient if differential treatment - adequate to the problems of the patient - is offered.
Zusammenfassung. Theoretischer Hintergrund. Suizid ist weltweit die zweithäufigste Todesursache bei jungen Menschen. Es scheint sinnvoll, speziell auf diese Zielgruppe zugeschnittene Angebote anzubieten. Fragestellung. Ziel dieser Untersuchung ist, den von den Nutzenden wahrgenommenen Beratungserfolg eines Online-Peer-Suizidpräventionsprogramms zu ermitteln und zu prüfen, welche Merkmale mit dem Beratungserfolg zusammenhängen. Methode. Daten eines Online-Fragebogens wurden ausgewertet ( N = 318). Ergebnisse. Bei zwei Drittel der Befragten verminderte sich die Suizidalität. Wichtige Variablen zur Vorhersage der Veränderung der Suizidalität waren der Eindruck der Ratsuchenden, dass die Peers die persönlichen Anliegen verstehen und die Anregung zu einer Verhaltensänderung. Das Peerverhalten trug signifikant zur Vorhersage der subjektiven Nützlichkeit und der allgemeinen Situation bei. Schlussfolgerungen. Die Studie liefert positive Ergebnisse hinsichtlich der Evaluierung der Online-Suizidpräventions-Beratung. Da das Verstanden fühlen und das Peerverhalten im Zusammenhang mit dem Beratungserfolg steht, könnte der Peer-Ansatz und die damit einhergehende positive Beziehungserfahrung wichtig sein.
This study examines whether animal-assisted therapy for overweight children who take part in an established multimodal intervention programme provides an additional positive effect. It examines the question of whether the Body Mass Index of obese children (n = 15) can be lowered more using animal-assisted therapy (AAT) than in a control group (n = 14). The effects of AAT on motoric performance and psychosocial aspects were also investigated. Taking into account differences at the start of the animal-assisted intervention between the two groups, a decrease in standardised BMI for the intervention
group after six months of intervention was observed. Animal-assisted therapy also had a positive effect on the sport-related motoric variables of endurance and dexterity. External assessments by the parents of the intervention group showed significant improvements with regard to „antisocial behaviour“, „aggressive behaviour“ and „social problems“. Taken together, the results show that while animal-assisted therapy is an effective therapy component it does not appear to be an „easy“ therapeutic approach for reducing overweight in children. Animal-assisted therapy requires careful planning regarding the persons involved and the participating children.
Quality management in psychotherapy is necessary to reduce the number of therapies that draw a negative result. In therapy with adults, there are already existing models which are internationally verified. Also measuring equipment for the field of adult therapy is derived from this. These models and the existing measuring equipment were adapted in this study on the area of children and adolecents' treatment and tested in a field study on the possibility of their usage in outpatient children and adolescents psychotherapy. It turned out that the procedure in the outpatient area is easily usable. The active factors in the field of adults seem not to be adequate in children and adolescents. The results of this study show that the instruments to control process quality do not lead to the expected improvement of the treatment compared to the control group without the use of the newly developed measuring instruments.
Background:
Clinical studies have explored the relationship between toothbrushing and the development of recession, but relevant recession data for the multi-directional power toothbrush (PT) are lacking. The aim of this study was to evaluate the effect of brushing with either a multi-directional PT or an ADA reference manual toothbrush (MT) on pre-existing mid-buccal gingival recession (Pre-GR) over 12 months.
Methods:
This was a 12-month, prospective, single-blind, parallel-group, randomized controlled clinical study. Healthy participants without periodontitis but with at least 2 teeth showing Pre-GR ≥2mm were randomized to a group either brushing with a MT or a PT. The primary outcome parameter was the change at sites with Pre-GR ≥2mm. All clinically-based recession measurements were performed by one calibrated examiner at baseline, 6 and 12 months. Secondary outcomes were changes of recession at all mid-buccal sites (with or without pre-GR), changes in the percentage of recession sites demonstrating a change of ≥1mm as well as changes in pocket probing depths.
Results:
107 participants completed the study (PT: 55; MT: 52). Over the 12-month study period the mean recession at sites with Pre-GR ≥2mm decreased significantly from 2.2 to 2.1 mm in both groups (p<0.05). The extent of the recession parameters did not differ between MT and PT groups at any time point. Recession evaluated clinically and on stone casts was well correlated.
Conclusion:
Neither the PT nor the MT led to an increase in Pre-GR during 12 months of daily use.
Anhand von 80 Therapievideos und einer Fragebogenbatterie werden in der korrelativen Videographiestudie folgende Fragen beantwortet: Erstens - wie stellen sich vollstationäre psychosomatische Patienten während der ersten Therapiesitzung gegenüber ihrem Therapeuten dar? Hierbei zeigt sich, dass die 80 Psychotherapiepatienten dieser Studie im Durchschnitt am häufigsten ein guter Klient sein sowie eine gute Beziehung zum Therapeuten schaffen wollen und Interesse und Bereitschaft zeigen, an der Lösung ihrer Probleme zu arbeiten. Zweitens welche Variablen beeinflussen die patientische Selbstdarstellung in der ersten Psychotherapiesitzung? Bei der Beantwortung dieser Frage wird ein besonders Augenmerk auf Zusammenhänge der Selbstdarstellung der Patienten im Erstgespräch zu den Variablenbereichen Therapiemotivation, Depression und Persönlichkeit gelegt. Bezüglich des Variablenbereichs Therapiemotivation wird auch ein eventueller Rentenwunsch der Patienten und dessen Zusammenhang zur Selbstdarstellung untersucht. Die Ergebnisse zeigen, dass am meisten motivations- und anreizbezogene Variablen, insbesondere ein Rentenwunsch, und vergleichbar geringer störungsbezogene, soziodemographische und persönlichkeitsbezogene Merkmale eine Rolle bezüglich der patientischen Selbstdarstellung im Aufnahmegespräch spielen. So klagen z.B. Patienten mit Rentenwunsch im Aufnahmegespräch mehr und präsentieren sich hilfesuchender gegenüber ihrem Therapeuten als Patienten ohne Rentenwunsch. Eine dritte Frage lautet: Kann die Art und Weise, wie sich stationäre Psychotherapiepatienten in der ersten Sitzung gegenüber ihrem Therapeuten präsentieren, den späteren Therapieerfolg dieser Patienten vorhersagen? Es zeigt sich, dass der Therapieerfolg durch die Selbstdarstellung nicht vorhergesagt werden kann. Im Rahmen dieser Arbeit umfasst Selbstdarstellung alle markanten gesprochenen und anderweitig gezeigten Verhaltensweisen und Darstellungsformen, mit denen Menschen ihr Bild von sich vermitteln.
To treat people with occupational contact dermatitis, the German Accident Prevention and Insurance Association in the Health and Welfare Services offers 2-day individual prevention (IP) seminars.
We investigated whether there are short-term and medium-term changes in proximal (e.g. behaviour) and distal (e.g. symptoms) outcomes after an IP seminar, whether changes in proximal outcomes are associated with changes in distal outcomes, and whether subgroups can be identified that benefit in particular.
In a prospective study, 502 participants of 85 IP courses completed the health education impact questionnaire (heiQ™) and skin symptom questionnaire (Skindex-29) at the start of the course, immediately thereafter, and after 6 months. Change was assessed according to standardized effect size. Regression techniques were used to analyse associations between proximal and distal outcomes.
After 6 months, participants showed improved self-management skills and preventive behaviour, and less fear of job loss, disease-related symptoms, and emotional distress. Significant associations between proximal and distal outcomes were found. Participants who felt more limited by their skin disease showed greater effects.
The results are consistent with the assumption that IP courses provide a range of benefits for people with occupational contact dermatitis. Changes in distal outcomes may be influenced by changes in proximal outcomes.
© 2015 John Wiley & Sons A/S. Published by John Wiley & Sons Ltd.
Background and objective:The present study compares patient characteristics and treatment outcome of 138 patients from a day-care psychosomatic rehabilitation and 540 patients from an inpatient psychosomatic rehabilitation. It was analyzed if patients differ in demographic-, clinical- and socio-medical characteristics, if there are different treatment effects between the 2 types of rehabilitations and if other patient characteristics can predict success of treatment. Methods:Data was assessed twice (at intake and at discharge of rehabilitation) by self-reported surveys and external assessment of clinical discharge reports. Success of treatment was measured by changes in somatic and mental complaints, psychological well-being, work motivation, subjective forecast of employment, and externally estimated ability to work. Results:Results indicate that day-care patients were higher burdened than inpatients at intake. They showed more complaints, had a higher number of days of incapacity and a higher chance of work difficulties. Furthermore, the prevalence of a migration background was higher. Both kinds of rehabilitation were successful in reducing somatic and mental complaints and in increasing psychological well-being. The "Würzburger Screening", which measures the subjective forecast of employment and occupational stress, turned out to be the best predictor for success of treatment. Conclusion:Both day-care- and inpatient rehabilitation were mostly successful, even if day-care patients were more burdened at intake. Therefore day-care rehabilitation can be seen as a good alternative to inpatient rehabilitation.
© Georg Thieme Verlag KG Stuttgart · New York.
Nach Ablauf der 6-jährigen Förderdauer des Verbundforschungsprogramms „Rehabilitationswissenschaften” soll ein Resümee der Arbeit und der Erfahrungen in den Methodenprojekten vorgenommen werden. Haben die insgesamt acht Forschungsverbünde auch inhaltlich ein unterschiedliches Forschungsprofil aufgewiesen, so zeichneten sie sich durch zwei Einrichtungen aus, die in allen acht Verbünden mit vergleichbarer Aufgabenstruktur vorhanden waren. Dies sind zum einen die Geschäftsstelle eines jeden Verbundes und zum anderen das sog. Metho-denprojekt, durch das die forschungsmetho-dische Infrastruktur und Qualität der Projekte sichergestellt werden sollten. Im Beitrag soll die geleistete methodische Beratung kurz skizziert, die unterschiedlichen Aufgaben erläutert sowie ein Plädoyer für die Notwendigkeit einer wissenschaftlich fundierten Methodenberatung in der Verbundforschung geführt werden. Abschließend werden einige Aspekte diskutiert, die für die methodischen Standards zukünftiger Rehabilitationsforschung berücksichtigt werden sollten.
In internationalen Studien scheint die Effektivität multimodaler Behandlungsprogramme bei Patienten mit chronischen Rückenschmerzen im Rahmen stationärer Behandlung gesichert [1]. Jedoch fallen in den in Deutschland durchgeführten Studien Effekte zu den mittel- bis langfristigen Katamnesen nur mäßig aus [2] [3]. Faktoren, die dafür verantwortlich gemacht werden, sind zum einen die fortgeschrittene Chronifizierung der Patienten sowie nicht behandelte psychische Belastungen und zum anderen berufliche Problemlagen. Im Rahmen des IopKo-Projektes wurde eine Reihe von Maßnahmen zur nachhaltigen Förderung der Krankheitsbewältigung und der beruflichen Wiedereingliederung von Patienten mit Erkrankungen des Bewegungsapparates entwickelt und evaluiert. Diese Maßnahmen umfassen: (1) intensive multiprofessionelle Eingangsdiagnostik und beschleunigte Zuweisung zu Psychologen und Sozialdienst bei beruflichen und psychischen Problemen; (2) die Bildung (hinsichtlich der Teilhabestörung) homogener Patientengruppen auf der Basis einer multiprofessionellen Diagnostik; (3) differenzielle Behandlungsangebote, darunter ein multimodales Therapieprogramm (Rückenfit) für Patienten mit einem hohen Chronifizierungsrisiko und psychischen Problemen; (4) interaktive Schulungsmodule zur Praxis der Leistungsbeurteilung, des Rentenrechtes und zu Maßnahmen der beruflichen Rehabilitation (Teilhabe am Arbeitsleben) und (5) arbeitsplatznahe rehabilitative Trainingsmodule. Methode: Alle Maßnahmen, die im Rahmen des Konzeptes eingeführt wurden, waren Gegenstand einer kontrollierten prospektiven Bewertungsstudie, in der die Outcomes der Standard-Reha mit denen des integrierten Konzeptes verglichen wurden. Teilnehmer der Studie waren 482 Patienten der Klinik Münsterland, von denen 307 der Studiengruppe und 176 der Kontrollgruppe zugeordnet wurden. Die Teilnehmer der Studiengruppe erhielten neben einem individuellen Behandlungsprogramm die oben genannten Maßnahmen, während diese zur Erhebung der Kontrollgruppe ausgesetzt wurden. Die Daten für die Kontroll- und Studiengruppe wurden alternierend in Zeitblöcken von drei Monaten erhoben. Ergebnisse für die Gesamtstichprobe: Einschränkungen der Funktion, Schmerzen und psychische Belastungen hatten sich zu Ende des stationären Aufenthaltes und auch 10 Monate danach in der Studiengruppe stärker reduziert als in der Vergleichsgruppe. Die Arbeitsunfähigkeitstage hatten sich 10 Monate nach Entlassung im Vergleich zu einem analogen Zeitraum vor der Reha um 75 % reduziert. In Bezug auf die genannten Parameter ergaben sich in der Studiengruppe moderate bis starke Effekte, die die der Kontrollgruppe übertrafen. Ergebnisse für die Teilstichprobe von Patienten mit hohem Chronifizierungsrisiko: Patienten einer Teilstichprobe mit fortgeschrittener Schmerzchronifizierung oder hohem Chronifizierungsrisiko, die das Therapieprogramm „Rückenfit” durchliefen, verbesserten sich ebenfalls stärker als Patienten mit vergleichbaren Einschränkungen, die ein Standard-Reha-Programm erhielten. Auch für diese Teilstichprobe fanden wir überlegene moderate bis starke Effekte im Hinblick auf den funktionalen und somatischen Status. Stärker als Patienten der Gesamtstichprobe verbesserten sich diese Patienten im Hinblick auf die Depressivität und die psychische Belastung. In Bezug auf die Schmerzbelastung verbesserten sich beide Gruppen, Kontroll- und Studiengruppe, erheblich, aber ohne Unterschied. Schlussfolgerung: In dieser Studie konnte nachgewiesen werden, dass stationäre orthopädische Rehabilitation in einem konsequent interdisziplinären Setting, mit einer multimodalen Therapie und einem Fokus auf Aktivierung und Motivierung nachhaltige positive Effekte sowohl in Bezug auf die von Patienten wahrgenommene körperliche und psychische Besserung als auch auf ökonomische Parameter wie die Dauer der Arbeitsunfähigkeit oder die Inanspruchnahme medizinischer Leistungen haben kann. Wir interpretieren die Effekte (1) als Ergebnis der multiprofessionellen Diagnostik und Zuweisung, die hilft, die inhomogene Gruppe der Patienten mit unspezifischen Kreuzschmerzen in spezifischere und damit homogenere Behandlungsgruppen zu unterteilen, (2) als Ergebnis der gesteigerten Therapiemotivation durch die geschlossene Gruppe des multimodalen Programms, (3) als Ergebnis der Information, Aufklärung und intensivierten Beratung bei beruflichen Problemen, (4) als Ergebnis des arbeitsplatznahen Trainings sowie (5) als Ergebnis der schnelleren Erkennung und effizienteren Behandlung psychischer Belastungen. Die Ergebnisse zeigen die Bedeutung der stationären Rehabilitation, die effektiv sein kann, wenn die zu den Problemen des Patienten passenden differenziellen Behandlungsangebote gemacht werden.
Die vorliegende Arbeit verfolgt die Zielsetzung, das Wissen über soziallagespezifische Empowermentprozesse anhand von katamnestischen Daten aus Vorsorge- und Rehabilitationseinrichtungen für Mütter und ihre Kinder zu fundieren. Im Mittelpunkt der Untersuchung steht die Beantwortung folgender Fragen: 1. Wie entwickelt sich die psychische Gesundheit in Abhängigkeit von der sozialen Lage nach der Intervention? 2. welche soziallagespezifischen Empowermentprozesse bilden sich poststationär ab? und 3. welche Bedeutung kommt Empowerment für die nachhaltige Verbesserung der psychischen Symptombelastung zu? Die Analyse basiert auf Daten von 6 094 Patientinnen aus 39 Vorsorge- und Rehabilitationseinrichtungen, die bis zu einem Jahr nach der Intervention erhoben wurden. Für die Evaluation von Empowerment wurden zwei Itembatterien entwickelt, die sich auf kleinräumige Veränderungen der Lebenssituation (‚Verhältnis-Empowerment’) sowie auf Optimierungen von gesundheitsbezogenen Verhaltensweisen, Einstellungen und Kompetenzen (‚Verhaltens-Empowerment’) beziehen. Veränderungen im Gesundheitszustand wurden hinsichtlich der Dimension ‚psychische Symptombelastung’ (SCL-K-9) erfasst. Die soziale Schichtzugehörigkeit wurde clusteranalytisch bestimmt, die soziallagespezifische Gesundheitsrelevanz von Empowerment regressionsanalytisch ermittelt. Die gesundheitlich am stärksten belasteten Mütter der unteren sozialen Lage profitieren kurzfristig am meisten von der Intervention, jedoch gehen die Therapieeffekte in dieser sowie in der mittleren Soziallage nach der Rückkehr in den häuslichen Alltag überdurchschnittlich stark zurück. Insbesondere für die Mütter der unteren und mittleren Soziallage erwies sich das ‚Verhaltens-Empowerment’ als bedeutsam für die nachhaltige Verbesserung des psychischen Gesundheitszustandes, allerdings sind die Empowermenterfolge hier im Vergleich zu den Müttern der oberen Soziallage vor allem in der Katamnese nach sechs Monaten deutlich geringer. In der Katamnese nach zwölf Monaten ist ein deutlicher Rückgang in der Bedeutung von Empowerment zu verzeichnen, wobei die relative Bedeutung des Verhältnis-Empowerments in der langfristigen Perspektive zunimmt. Die theoretische Evidenz von Empowerment konnte empirisch bestätigt werden. Der Umstand, dass die Gesundheitseffekte zwölf Monate nach der Intervention deutlich zurückgehen, weist auf die Notwendigkeit von unterstützenden poststationären therapeutischen Maßnahmen hin. Die ermittelten geringen Unterschiede zwischen den Empowermenterfolgen der unteren und mittleren Soziallage legen nahe, dass sozialschichtspezifische Analysen im vorliegenden Fall nur begrenzt aussagekräftig sind.
Background
The role of placebos is often misunderstood, leading both to overvaluation and to inappropriate disdain. The effect of a placebo that contains no pharmacologically active substance is often confused with the effect of administration by a physician. The aim of this article is to review the current data on placebos, evaluate these data critically, and provide a well-founded and understandable explanation of the effects that placebos do and do not possess.Methods
Selective literature review.ResultsRecent studies employing modern imaging techniques have provided objective correlates of the effect of placebo administration for certain indications. A recent paper even suggested a genetic basis for it. Two main mechanisms underlie the effect of placebo administration: conditioned reflexes, which are subconscious, and the patient's expectations, which are conscious. Further factors include the physician's personality and the setting in which the treatment takes place.Conclusions
The mechanisms of action of placebo administration, with which positive therapeutic effects can be achieved with little effort, should be consciously exploited by physicians when giving their patients pharmacologically active medications as well.
The analytical quality of self-monitoring of blood glucose (SMBG) is not satisfactory, and the need for standardized control routines for SMBG has been underlined. The objective of this study was to investigate how to perform quality assurance of glucometers for SMBG in a physician's office or at an outpatient clinic, using a split sample design. One hundred eighteen diabetes patients at 20 offices of general practitioners and 1 hospital outpatient clinic participated in the study. Each patient met for 1 consultation where they performed glucose measurements under observation of a biomedical laboratory scientist using their own meter. Glucose measurements were also performed once on an office method and on a conventional reference method. The patient's result was compared with the office method and was assessed as acceptable if the result was within the quality specifications recommended in International Organization for Standardization 15197. If the result was outside the limits, the patient was educated in the correct use of their instrument before new measurements were performed. Thirteen percent of the patients' results were outside the acceptable limits. After the second measurement, 3 patients still obtained unacceptable results, but reasonable explanations for the poor results were found. A total of 19% of the patients made some kind of user errors, but only 5% of the measurements were related to poor results. The present quality assurance procedure can be used when caring for patients using SMBG. The control program implies that the office can document acceptable analytical quality of their method and good knowledge of SMBG systems.
The role of placebos is often misunderstood, leading both to overvaluation and to inappropriate disdain. The effect of a placebo that contains no pharmacologically active substance is often confused with the effect of administration by a physician. The aim of this article is to review the current data on placebos, evaluate these data critically, and provide a well-founded and understandable explanation of the effects that placebos do and do not possess.
Selective literature review.
Recent studies employing modern imaging techniques have provided objective correlates of the effect of placebo administration for certain indications. A recent paper even suggested a genetic basis for it. Two main mechanisms underlie the effect of placebo administration: conditioned reflexes, which are subconscious, and the patient's expectations, which are conscious. Further factors include the physician's personality and the setting in which the treatment takes place.
The mechanisms of action of placebo administration, with which positive therapeutic effects can be achieved with little effort, should be consciously exploited by physicians when giving their patients pharmacologically active medications as well.
On completion of the funding period of six years within the federal programme for research in medical rehabilitation, an evaluative summary is given of the work performed and the experiences made in the cross-sectional projects offering methodological support. While specific research profiles had prevailed in each of the eight regional research networks, two institutions had been implemented with similar task assignments within all networks; i. e., (i) a central office, and (ii) a center for methodological support (CMS) had been available in order to enhance methodological research infrastructure and research quality of the clinical projects. This article outlines the support offered as well as further tasks of the CMS. Further, it is argued that organized, scientifically based methodological support and consultation should be part of any research network. Finally, a number of important aspects are discussed which should be taken into account to enhance research quality in medical rehabilitation in the future.
Examines regression effects in longitudinal sequences of observations by formulating expectations (Es) for later observations conditioned on an initial selection score value. The Es are developed for several variations of classical test theory and autocorrelation models. Es based on the general concept of regression are seen not only to depart from those depicted in the psychometric lore but to vary considerably from one underlying model to another, particularly when moving from the 2-occasion to a multiple-occasion measurement framework. Regression toward the mean is not a ubiquitous phenomenon, nor does it always continue across occasions. It is necessary to specify the characteristics of model assumptions to understand the when, how, and extent of regression toward the mean. Past interpretations have been incomplete and to an extent incorrect because they focused largely on a limited circumstance: 2-occasions of measurement and simplexlike correlation matrices. (23 ref) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Examines procedures previously recommended by various authors for the estimation of "change" scores, "residual," or "basefree" measures of change, and other kinds of difference scores. A procedure proposed by F. M. Lord is extended to obtain more precise estimates, and an alternative to the L. R. Tucker, F. Damarin, and S. A. Messick (see 41:3) procedure is offered. A consideration of the purposes for which change measures have been sought in the past leads to a series of recommended procedures which solve research and personnel-decision problems without estimation of change scores for individuals. (22 ref.) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Explicates the fundamental nature of regression toward the mean, which is frequently misunderstood by developmental researchers. While errors of measurement are commonly assumed to be the sole source of regression effects, the latter also are obtained with errorless measures. The conditions under which regression phenomena can appear are first clearly defined. Next, an explanation of regression effects is presented which applies both when variables contain errors of measurement and when they are errorless. The analysis focuses on cause and effect relationships of psychologically meaningful variables. Finally, the implications for interpreting regression effects in developmental research are illustrated with several empirical examples. (21 ref.) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
Changes in the loss of working hours (LWH) after a course of spa treatment because of illness has often been used as a parameter for the success of that kind of therapy. The fitness of the use of this method is examined on the basis of comparison with a control group, and the results are critically tested in accordance with statistical basic principles. First there is an investigation of the notion of, and the mathematical statistical working-out of the LWH. Valid comparisons between 2 periods of time are possible only when the LWHs are treated, not as isolated cases, but are comprehended as the total values for one person. In this, nil values of the LWH must be included in the calculation. In the comparison of the total sums of patients between 2 successive time periods, structural regroupings occur, as in the overwhelming majority of persons these values for LWH change in the subsequent period of time. In a collectivity of women working in a job it is shown that these changes occur in dependence on the initial value. After nil values of the LWH in the previous period of time there follow increases of the LWH; after values of about 3 weeks duration, on the average, similar times are to be found; after higher values there are always recorded in increasing degree, a falling-off of the LWH. The changes of these total patient values for the LWH are known as the 'epidemiological trend'. In a collectivity the sum of all changes is nil when the condition of the patients remains constant.
Zusammenfassung Verschiedene Strukturgleichungsmodelle zur Beschreibung und Vorhersage von Unterschieden in längsschnittlichen Veränderungen werden dargestellt. Für die Beschreibung längsschnittlicher Veränderung in einem Merkmal werden ein Ko-variationsmodell, ein trait/state-Modell, ein autoregressives Modell sowie ein Modell von Startwert und absoluter Merkmalsveränderung dargestellt. Im Kovariations-und im trait/state-Modell kann Merkmalsvarianz nur durch aktuell wirksame Faktoren entstehen; mit dem Wegfall der Wirkfaktoren verschwindet auch die Varianz im gemessenen Merkmal (Elastizität der Merkmalsausprägung). Im autoregressiven Modell wird dagegen die Möglichkeit der Selbstperpetuierung von Merkmalsvarianz eingeräumt (Merkmalsträgheit); im Modell von Startwert und absoluter Veränderung wird die Merkmalsstarrheit sogar fest ins Modell eingebaut, neu hinzukommende Wirkfaktoren führen zu zusätzlicher Merkmalsvarianz (Modell kumulierender Einflüsse). Bei den Vorhersagemodellen werden zunächst verschie-dene Varianten der direkten Vorhersage (ohne intervenierende latente Faktoren) dargestellt, die sich danach unterscheiden, ob das Modell die Schätzung zeitversetz-ter Effekte und autoregressiver Trägheitseffekte erlaubt. Anschließend werden verschiedene Versionen von Vorhersagemodellen spezifiziert, die für mindestens eine der beiden Variablen (Prädiktor und/oder Kriterium) eine Varianzzerlegung nach dem trait/state-Modell enthalten.
Explored whether a true sophomore slump (a significant decline in competitive performance during athletes' 2nd yr of competition following an outstanding 1st season) exists or whether it is a misconception based on a lack of understanding of regression toward the mean. Archival performance data of Major League baseball hitters (between 1960 and 1983) and pitchers (between 1945 and 1983) who had outstanding rookie seasons were examined. Results provided equivocal support for the 2 proposed explanations. Consistent with the sophomore slump view, 1 performance measure, home runs, demonstrated the expected pattern at a significant level. However, batting average findings were more consistent with the regression toward the mean interpretation. (French, Spanish, German & Italian abstracts) (PsycINFO Database Record (c) 2012 APA, all rights reserved)
chapter is concerned with methods for the analysis of longitudinal data
seeks to convey "right thinking" about longitudinal research
heroes of this chapter are statistical models for collections of individual growth (learning) curves
myths / indicate some of the beliefs that have impeded doing good longitudinal research (PsycINFO Database Record (c) 2012 APA, all rights reserved)
We have previously shown that regression towards the mean occurs whenever we select an extreme group based on one variable and then measure another variable for that group (4 June, p 1499).1 The second group mean will be closer to the mean for all subjects than is the first, and the weaker the correlation between the two variables the bigger the effect will be. Regression towards the mean happens in many types of study. The study of heredity1 is just one. Once one becomes aware of the regression effect it seems to be everywhere. The following are just a few examples.Treatment to reduce high levels of a measurement - In clinical practice there are many measurements, such as weight, serum cholesterol concentration, or blood pressure, for which particularly high or low values are signs of underlying disease or risk factors for disease. People with extreme values of the measurement, such as high blood pressure, may be treated to bring their values closer to the mean. If they are measured again we will observe that the mean of the extreme group is now closer to the mean of the whole population - that is, it is reduced. This should not be interpreted as showing the effect of the treatment. Even if subjects are not treated the mean blood pressure will go down, owing to regression towards the mean. The first and second measurement will have correlation r<l because of the inevitable measurement error and biological variation. The difference between the second mean for the subgroup and the population mean will be approximately r times the difference between the first mean and the population mean. We need to separate any genuine reductions due to treatment from the effect of regression towards the mean. This is best done by using a randomised control group, but it can be estimated directly.2Relating change to initial value - We may be interested in the relation between the initial value of a measurement and the change in that quantity over time. In antihypertensive drug trials, for example, it may be postulated that the drug's effectiveness would be different (usually greater) for patients with more severe hypertension. This is a reasonable question, but, unfortunately, the regression towards the mean will be greater for the patients with the highest initial blood pressures, so that we would expect to observe the postulated effect even in untreated patients.3Assessing the appropriateness of clinical decisions - Clinical decisions are sometimes assessed by asking a review panel to read case notes and decide whether they agree with the decision made. Because agreement between observers is seldom perfect the panel is sure to conclude that some decisions are “wrong.” For example, Barrett et al reviewed cases of women who had had a caesarean section because of fetal distress.4 The percentage agreement between pairs of observers in the panel varied from 60% to 82.5%. They judged a caesarean section to be “appropriate” if at least four of the five observers thought a caesarean should have been done. Because there was poor agreement among the panel, judgments by panel members and the actual obstetricians doing the sections must also be poorly related and not all caesareans will be deemed appropriate by the panel. The authors concluded that 30% of all caesarean sections for fetal distress were unnecessary, but what the study actually showed was that decisions about whether women should have emergency surgery for fetal distress are difficult and that obstetricians do not always agree.5Comparison of two methods of measurement - When comparing two methods of measuring the same quantity researchers are sometimes tempted to regress one method on the other. The fallacious argument is that if the methods agree the slope should be 1. Because of the effect of regression towards the mean we expect the slope to be less than 1 even if the two methods agree closely. For example, in two similar studies self reported weight was obtained from a group of subjects, and the subjects were then weighed.6,7 Regression analysis was done, with reported weight as the outcome variable and measured weight as the predictor variable. The regression slope was less than 1 in each study. According to the regression equation, the mean reported weight of heavy subjects was less than their mean measured weight, and the mean reported weight of light subjects was greater than their mean measured weight. We have a finding which allows a simple and attractive, but misleading, interpretation: those who are overweight tend to underestimate their weights and those who are excessively thin tend to overestimate their weights. In fact we would expect to find a slope less than 1, as a result of regression towards the mean. If self reported and measured weight were equaly good measures of the subject's true weight then the slope of the regression of reported weight on measured weight will be less than 1. But the slope of the regression of measured weight on reported weight will also be less than 1. Now we have the oppostive conclusion: people who are heavy have overestimated their weights and people who are light have underestimated theirs. Elsewhere we describe a better approach to such data.8Publication bias - Rousseeuw notes that referees for papers submitted for publication do not always agree which papers should be accepted.9 Because referees' judgments of the quality of papers are therefore made with error, they cannot be perfectly correlated with any measure of the true quality of the paper. Thus when an editor accepts the “best” papers for publication the average quality of these will be less than the editor thinks, and the average quality of those rejected will be higher than the editor thinks. Next time you are turned down by the BMJ do not be too despondent. It could be just another example of regression towards the mean.References1.↵Bland JM, Altman DG. Regression towards the mean. BMJ 1994;308:1499.OpenUrlFREE Full Text2.↵Davis CE. The effect of regression to the mean in epidemiologic and clinical studies. Am J Epidemiol 1976;104:493-8.OpenUrlFREE Full Text3.↵Hayes RJ. Methods for assessing whether change depends on initial value. Statistics in Medicine 1988;7:915-27.OpenUrlMedlineWeb of Science4.↵Barrett JFR, Jarvis GJ, Macdonald HN, Buchan PC, Tyrrell SN, Lilford RJ. Inconsistencies in clinical decision making in obstetrics. Lancet 1990;336:549-51.OpenUrlCrossRefMedlineWeb of Science5.↵Esmail A, Bland M. Caesarian section for fetal distress. Lancet 1990;336:819.OpenUrl6.↵Kuskowska-Wolk A, Karlsson P, Stolt M, Rossner S. The predictive value of body mass index based on reported weight and height. Int J Obesity 1989;13:441-3.OpenUrlMedlineWeb of Science7.↵Schilchting P, Hoilund-Carlsen Quaade F. Comparison of self reported height and weight with controlled height and weight in women and men. Int J Obesity 1981;5:67-76.OpenUrlMedlineWeb of Science8.↵Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet 1986;i:307-10.9.↵Rousseeuw PJ. Why the wrong papers get published. Chance 1991;4:41-3.OpenUrl
Abstract High measurement values often show on average a spontaneous decrease when remeasured under stationary study conditions. This effect is known as “regression to the mean”, a phenomenon widely met in biomedical research. In this paper a general formula is derived, which shows that this effect should be better called “regression to the mode”. Further it is shown that this effect may depend on the time-spacing of repeated measurements in a stationary population.
Investigators must take great care in studying the temporal association between biological and clinical data. By selecting patients on the basis of deviant levels of a biological parameter, it often is impossible to separate the effect of regression toward the mean from a temporal association of clinical interest. Experimental strategies that focus exclusively on subjects with extreme values are particularly prone to misinterpreting the effects of regression toward the mean as a temporal change in the variable of interest. We cite specific examples of two previous studies of the relationship between the dexamethasone suppression test (DST) and clinical response in depression. Contrary to the conclusions of the investigators, normalization of the DST had no relation to clinical response in the first 5 weeks of treatment. The decrease of post-dexamethasone cortisol levels occurred regardless of clinical response, most likely due to the effect of regression toward the mean.
A general maximum likelihood approach for estimating the effects of treatments applied to samples subject to regression to the mean is outlined. Models may be specified in terms of three factors: whether the treatment effect is multiplicative or additive, whether the treatment group is above or below some truncation point and the type of sample involved. The way in which solutions may be obtained for all 16 models so defined is described.
The notion of “regression to the mean” is widely misunderstood. This paper explains the concept in simple terms and shows how it arises in studies of mental and physical development.
Der Begriff “Regression auf den Mittelwert” wird häufig mißverstanden. Diese Arbeit erklärt das Konzept in einfacher Weise und zeigt, wie es in Studien der geistigen und körperlichen Entwicklung entsteht.
La notion de “régression à la moyenne” est largement mal comprise. Ce travail explique le concept en termes simples et montre comment il survient dans des études de développement mental et physique.
When analyzing data from situations in which before and after treatment measurements have been obtained, without a parallel control group, it is important that the investigator be aware of possible changes in the observed variable due to regression toward the mean. This phenomenon can be misleading and is sometimes overlooked in the interpretation of data. This paper demonstrates the importance of a control group by distinguishing between the true treatment effect and the regression effect for the case of bivariate normal distribution, truncated with regard to the predictor or "before treatment" measurement. The regression effect is noted to be especially significant when the correlation between the before and after treatment measurement is small. A method for estimating the treatment and regression effects is presented and applied to data from an actual study.
In many randomised trials researchers measure a continuous variable at baseline and again as an outcome assessed at follow up. Baseline measurements are common in trials of chronic conditions where researchers want to see whether a treatment can reduce pre-existing levels of pain, anxiety, hypertension, and the like.
Statistical comparisons in such trials can be made in several ways. Comparison of follow up (post-treatment) scores will give a result such as “at the end of the trial, mean pain scores were 15 mm (95% confidence interval 10 to 20 mm) lower in the treatment group.” Alternatively a change score can be calculated by subtracting the follow up score from the baseline score, leading to a statement such as “pain reductions were 20 mm (16 to 24 mm) greater on treatment than control.” If the average baseline scores are the same in each group the estimated treatment effect will be the same using these two simple approaches. If the treatment is effective the statistical significance of the treatment effect by the two methods will depend on the correlation between baseline and follow up scores. If the correlation is low using the change score will …
This paper gives a critical view on "Goal Oriented Measurement" (ZOE, Zielorientierte Ergebnismessung) with the IRES (Indicators of Rehabilitation Status) questionnaire. Gerdes (1998) proposed this method for individualizing outcome measurement in rehabilitation research. The rationale of Goal Oriented Measurement is described and differentiated from Goal Attainment Scaling and Patient Preferences. Subsequently, important shortcomings of Goal Oriented Measurement are pointed out: insufficient validity with regard to selecting clinically relevant problems, possible regression artifacts, unsuitably aggregated scores, possible reactivity, possible changes of individual problems, possible deteriorations in problems not initially selected, and plurality as well as small size of subsamples. Overall, Goal Oriented Measurement has to undergo substantial improvement.
Bewirken Rehamaßnahmen eine Abnahme der Arbeitsunfä− higkeit? Eine Fall−Kontroll−Studie
- N Gerdes
Gerdes N. Bewirken Rehamaßnahmen eine Abnahme der Arbeitsunfä−
higkeit? Eine Fall−Kontroll−Studie. Stuttgart: Thieme, 1993
- J A Ingelfinger
- F Mosteller
- L Thibodeau
- J H Ware
Ingelfinger JA, Mosteller F, Thibodeau L, Ware JH. Biostatistics in clini−
cal medicine. 3. Aufl. New York: McGraw−Hill, 1994
Ursache vieler Denkfehler und Fehl− beurteilungen
- K Fiedler
- Die Regressionsfalle
Fiedler K. Die Regressionsfalle. Ursache vieler Denkfehler und Fehl−
beurteilungen. Ruperto Carola, Forschungsmagazin der Universität
Heidelberg 2000; (2): 16 ±22 ± verfügbar im Internet unter: www.uni−
heidelberg.de/presse/ruca/ruca2_2000/falle.html (aufgerufen 5.4.
2005)
Statistisch−methodische Anmerkungen zum epidemio− logischen Arbeitsunfähigkeitstrend
- M Stallmann
Stallmann M. Statistisch−methodische Anmerkungen zum epidemio−
logischen Arbeitsunfähigkeitstrend. Gesundheitswesen 1996; 58:
303 ± 313
Myths about longitudinal research Methodological issues in aging research
- D Rogosa
Rogosa D. Myths about longitudinal research. In: Schaie KW, Campbell
RT, Rawlings SC (eds): Methodological issues in aging research. New
York: Springer, 1988: 171 ± 209
Mythos und Wirklichkeit (methodenreport 4[2]) Jena: Friedrich−Schiller−Universität, Institut für Psychologie
- C Nachtigall
- U Suhl
- Der Regressionseffekt
16 Nachtigall C, Suhl U. Der Regressionseffekt. Mythos und Wirklichkeit
(methodenreport 4[2]). Jena: Friedrich−Schiller−Universität, Institut
für Psychologie, 2002 ± verfügbar im Internet unter: www2.uni−jena.
de/svw/metheval/materialien/reports/report_2002_02.pdf (aufgeru−
fen 5.4.2005)
Studying true change patterns via structural equation modeling Psychometric methodology
- T Raykov
- R Steyer
- Kf Wender
- Kf Widaman
Raykov T. Studying true change patterns via structural equation
modeling. In: Steyer R, Wender KF, Widaman KF (eds): Psychometric
methodology. Proceedings of the 7 th European Meeting of the Psycho−
metric Society in Trier. Stuttgart: Fischer, 1993: 430 ± 434
Zielorientierte Ergebnismessung (ZOE) mit dem IRES− Patientenfragebogen: Eine kritische Zwischenbilanz
- C Zwingmann
Zwingmann C. Zielorientierte Ergebnismessung (ZOE) mit dem IRES−
Patientenfragebogen: Eine kritische Zwischenbilanz. Rehabilitation
2003; 42: 226 ± 235
Methodological errors in medical research
- B Andersen
Andersen B. Methodological errors in medical research. Oxford: Black−
well, 1990
Regression to the mean
- C E Davis
Davis CE. Regression to the mean. In: Johnson NL, Kotz S (eds): Ency−
clopedia of statistical sciences, Vol. 8. New York: Wiley, 1986: 706 ±
708
Studying true change patterns via structural equation modeling
- T Raykov
Raykov T. Studying true change patterns via structural equation
modeling. In: Steyer R, Wender KF, Widaman KF (eds): Psychometric
methodology. Proceedings of the 7 th European Meeting of the Psycho−
metric Society in Trier. Stuttgart: Fischer, 1993: 430 ± 434
Fehlerquellen bei Kurerfolgsbeurteilungen mittels Arbeits− ausfallzeiten wegen Krankheit
- H Wagner
Wagner H. Fehlerquellen bei Kurerfolgsbeurteilungen mittels Arbeits−
ausfallzeiten wegen Krankheit. Zeitschrift für Physiotherapie 1977;
29: 313 ± 338
- M Wirtz
- C Deskriptive Nachtigall
- Statistik
Wirtz M, Nachtigall C. Deskriptive Statistik. Statistische Methoden für
Psychologen, Teil 1. 3. Aufl. Weinheim: Juventa, 2004