86
Views
2
CrossRef citations to date
0
Altmetric
Original Articles

An individual differences perspective on change in psychotherapy: The case of health care utilization

, &
Pages 690-705 | Received 07 Jun 2006, Published online: 17 Oct 2007
 

Abstract

The purpose of this study was to explore systematic interindividual variation in change of a number of health care utilization variables (HCUVs) during psychotherapy and identify patient characteristics associated with this variation. Three-wave panel data from 420 patients were analyzed with nonparametric latent class regression followed by chi-square interaction analyses among patient variables. For the various HCUVs, three to six classes were identified, with widely different patterns of change during treatment. Axis I diagnosis, chronicity, functional impairment, gender, and level of education were among the patient characteristics that differentiated the classes. It was concluded that main effects analyses seriously distort heterogeneity of change and that health care utilization, unless it is a specific therapeutic aim, may be irrelevant as an indication of outcome of psychotherapy.

Abstract

Die Perspektive individueller Unterschiede in der Veränderung der Psychotherapie: Nutzung der medizinischen VersorgungDer Grund dieser Untersuchung bestand darin systematische interindividuelle Unterschiede bezüglich der Veränderung der Nutzung der medizinischen Versorgung während einer Psychotherapie zu erforschen (HCUVs, Variablen der medizinischen Versorgungsnutzung) und die Charakteristika des Patienten, die damit einhergehen. Daten von 420 Patienten (3 Erhebungswellen) wurden mit einer nonparametrischen Latent-Class-Regression analysiert, gefolgt von einer Chi-Quadrat-Analyse der Patientenvariablen. Für die verschieden HCUs wurden drei bis sechs Klassen, mit sehr unterschiedlichen Mustern der Veränderung während der Behandlung, identifiziert. Achse 1 Diagnosen, Chronizität, funktionale Beeinträchtigungen, Geschlecht und Bildungsniveau waren die Patientencharakteristika, welche die Klassen differenzierten. Es wurde geschlussfolgert, dass eine Analyse der Haupteffekte die Heterogenität der Veränderung schwerwiegend verfälschen würde und dass die Nutzung der medizinischen Versorgung – obwohl es ein spezifisches therapeutisches Ziel ist – möglicherweise als ein Anzeichen für das Therapieergebnis (Outcome) irrelevant ist.

Abstract

Une perspective basée sur les différences individuelles au sujet du changement en psychothérapie: la question de l'utilisation des soins de santé

L'intention de cette étude était d'explorer la variation interindividuelle systématique du changement d'un certain nombre de variables d'utilisation des soins de santé (HCUVs) pendant la psychothérapie, et d'identifier les caractéristiques des patients associées avec cette variation. Les données de 420 patients en trois vagues étaient analysées par la régression non-paramétrique de classe latente suivie d'analyses chi-carrés de l'interaction entre les variables des patients. Pour les différentes HCUVs, trois à six classes ont été identifiées, avec des patterns de changement en cours de traitement largement différents.Le diagnostic Axe I, la chronicité, le handicap fonctionnel, le sexe, et le niveau d’éducation étaient parmi les caractéristiques des patients qui différenciaient les classes. Nous avons conclu que des analyses d'effet principal faussent sérieusement l'hétérogénéité du changement et que l'utilisation des soins de santé peut – sauf si c'est un but thérapeutique spécifique – être négligeable comme indicateur de l'efficacité d'une psychothérapie.

Abstract

Una perspectiva individual de las diferencias del cambio en psicoterapia. El caso de la utilización del servicio de salud.

El propósito de este estudio fue explorar las variaciones del cambio sistemático interindividual de un número de variables de utilización del servicio de salud (HCUVs) durante la psicoterapia e identificar las características del paciente asociadas con esta variación. Se analizaron tres paneles de datos (three-wave panel data) de cuatrocientos veinte pacientes con una regresión de clase latente no paramétrica (nonparametric latent class regression) seguida del chi-cuadrado de los análisis de la interacción entre las variables de los pacientes. Se identificaron tres a seis clases de los diversos HCUVs, con amplios patrones de cambio durante el tratamiento. Entre las características del paciente útiles para diferenciar las clases se contó con el diagnóstico en Eje I, la cronicidad, el deterioro funcional, el género y el nivel de educación. Se concluyó que los análisis de los efectos principales distorsionan seriamente la heterogeneidad del cambio y que la utilización de los servicios de salud, a menos que sea un objetivo terapéutico específico, puede ser irrelevante como indicación de resultado de una psicoterapia.

Abstract

A perspectiva das diferenças individuais da mudança em psicoterapia: O caso da utilização dos serviços de saúde.

O objectivo deste estudo foi explorar a variação sistemática interindividual na mudança de um número de variáveis da utilização dos serviços de saúde (VUSSs) durante a psicoterapia e identificar as características dos pacientes, associadas a esta variação. Foram analisadas três ondas de dados de 420 pacientes através de regressão de classe latente não paramétrica seguida de análises de interacção com qui-quadrado entre as variáveis dos pacientes. Para as várias VUSSs foram identificadas três a seis classes, com diferentes padrões de mudança durante o tratamento. Diagnóstico do Eixo I, cronicidade, comprometimento funcional, género e nível de educação foram as características dos pacientes que diferenciaram as classes. Conclui-se que as análises dos efeitos principais distorciam seriamente a heterogeneidade da mudança e que a utilização dos serviços, a não ser que seja uma ajuda terapêutica específica, pode ser irrelevante como indicação dos resultados da psicoterapia.

Abstract

Una prospettiva sulle differenze individuali sul cambiamento in psicoterapia: il caso dell'utilizzo della cura sanitaria

L'obiettivo di questo studio è l'esplorazione sistematica della variazione interindividuale del cambiamento delle variabili dell'utilizzo della cura sanitaria (UCS) durante la psicoterapia e l'identificazione delle caratteristiche dei pazienti associate con questo cambiamento. Dati di 420 pazienti con regressione a classi latenti non parametrica seguita da analisi delle interazioni chi-quadrato tra le variabili del paziente. Per i vari UCS sono stati identificati 3 delle 6 classi, con modelli di cambiamento ampiamente differenti durante il trattamento. Le caratteristiche del paziente che differenziavano le classi erano la diagnosi di asse I, la cronicità, l'invalidità funzionale, il sesso e il livello educativo. In conclusione, l'analisi degli effetti principali distorceva seriamente l'eterogeneità del cambiamento mentre l'UCS può essere irrilevante come indicatore di esito terapeutico, a meno che esso non sia un obiettivo terapeutico.

Notes

1. The too-small low-dose group was excluded to simplify the analysis of treatment type to a binary variable.

2. The model for counts is a log-linear Poisson regression,

ln(µ) ti =β01 * t i or

(µ)ti = exp(β0 + β1 * ti)

where µ is the population mean, β0 and β1 are intercept and slope estimates, respectively, and t i is the ordinal stage number. Thus, the natural logarithm of µ changes linearly with β1 per stage, whereas µ will change nonlinearly in most cases. Because the estimates may be difficult to comprehend at first glance, we also report the accumulated change in raw scores from the year before treatment to the third year after termination and, in cases where the trend is reasonably linear, the average change per stage.

3. The change pattern of each cluster is characterized in terms of the approximate relative levels, initially and finally, using the following descriptors: H = high; L = low; M = medium; VH = very high; VL = very low; Z=(close to) zero.

4. The covariates are ordered according to the strength of their association with the classification; the covariate with the strongest association is given first.

5. The paradoxical fact that the entire group developed in the increasing direction but all four latent classes developed in the decreasing direction deserves some reflection. The two-, three-, and five-class solutions—but not the four-class solution—had one class with increasing number of weeks. This is not likely a software artifact because another software yielded exactly the same pattern. Possibly patients with increasing number of weeks are divided among the four classes without substantially affecting the class average, although they contribute to increased within-class heterogeneity/error.

6. For category scales, the adjacent category logit model is specified. The β1 is a log-odds ratio, meaning that the odds of category j versus category j – 1 (j = 1, 2 …5 for a five-step scale) changes by exp(β1) per stage. The odds are assumed constant for all pairs of j and j – 1.

7. For a nominal variable, the variance explained represents the weighted average of separate squared multiple correlations for each category (class) taken as a dichotomous variable.

Log in via your institution

Log in to Taylor & Francis Online

PDF download + Online access

  • 48 hours access to article PDF & online version
  • Article PDF can be downloaded
  • Article PDF can be printed
USD 53.00 Add to cart

Issue Purchase

  • 30 days online access to complete issue
  • Article PDFs can be downloaded
  • Article PDFs can be printed
USD 200.00 Add to cart

* Local tax will be added as applicable

Related Research

People also read lists articles that other readers of this article have read.

Recommended articles lists articles that we recommend and is powered by our AI driven recommendation engine.

Cited by lists all citing articles based on Crossref citations.
Articles with the Crossref icon will open in a new tab.