What is oq 45

Necessary Necessary. Non-necessary Non-necessary. Are you OQ Measures user? Yes No. Thank You! Get A Quote. How do you want us to contact you?

Are you going to do a research study? Are you a student? How many full-time equivalent FTE clinicians do you intend to license? How did you hear about us? Check all that apply. Name of Colleague. Name of Conference. Specify Company or University. Name of Facebook Group.

In the outpatient setting, The following self-report and clinician-administered instruments were used for this analysis:. Basic Documentation. Important clinical information and sociodemographic characteristics were recorded on a form. These data were collected by the treating psychiatrists for inpatients and the treating psychotherapists for outpatients. Diagnoses based on this system were available for the inpatient samples and had been assigned by the treating psychiatrists.

These interviews, which were available only for sample 3, served in assessing Axis I and II disorders clinical syndromes and personality disorders according to the DSM-IV criteria. Trained psychologists not involved in treatment of the patients were engaged to carry out the two clinical interviews. This instrument gathers the clinician's view of the severity of psychopathology CGI-S and of the improvements from the initiation point of the treatment CGI-I.

The two aspects are each rated on a 7-point scale. Ratings with this instrument were available only for the inpatient samples and were provided by the treating psychiatrists. On this scale psychological, social, and occupational functioning are rated on a hypothetical continuum from severe mental illness 0 to mental health The scale structure of the original version is supported by confirmatory factor analysis [ 14 ]. The single scales in the German version exhibit good internal consistency [ 20 ].

This questionnaire is the short version of the SCLR [ 6 ]. Additional global indices of distress can be obtained, for example, the General Symptom Index GSI , which is the mean score of all items.

In the analysis carried out by Derogatis and Melisaratos [ 5 ] seven of the primary scales showed a clear convergence with their counterparts among the MMPI scales.

All scales in the German version exhibit satisfactory test-retest reliability [ 22 ]. Nonresponse is a common problem in surveys and can occur for single items or for the entire examination. Analyzing only data of patients that responded to every item on the self-report measures at admission as well as at discharge would lead to a substantial loss of information, especially for the analyses for the inpatient setting.

To include as many cases as possible, we defined different inclusion criteria for each analysis Figure 1. First, we defined the minimal number of answered items to calculate scale scores. In a previous unpublished study we tested through simulations the impact of incomplete item values in the calculation of scores for the OQ scales. In this study, we considered as evaluable only data records that fulfilled the rule mentioned above. Concerning the BSI, we analyzed only returned questionnaires with at least 40 valid item values [ 32 ].

We calculated the response rate on the basis of the number of returned questionnaires that fulfilled the above-mentioned criteria of completeness. Patients who completed a sufficient number of items at admission were included in the analysis of sensitivity to psychopathology.

Patients who additionally did the same at discharge were included in the analysis of sensitivity to change. For the analysis of intraindividual changes we applied Jacobson and Truax's method of calculating clinical significance [ 19 ], with which the proportions of improved and recovered cases can be calculated [ 2 ]. To determine these proportions, two parameters are needed: 1 a critical difference D , which allows identification of individual pre-post differences that are sufficiently large to be considered statistically significant reliable change and 2 a cut-off C , which distinguishes scores that are a variation of normal functioning from scores indicating a psychopathological state.

We additionally compared our parameter values with those published in other studies on the basis of samples from Germany. To compare sensitivity to change, measured with two different questionnaires within three different samples, the following analysis techniques were used: propensity score matching and linear mixed modeling.

Given that within the inpatient setting OQ and BSI data were collected in two different samples, we matched sample 1 and sample 2 using the propensity score technique proposed by Rosenbaum and Rubin [ 33 ]. It is primarily used in nonrandomized studies in order to build equivalent samples for causal inference, but it can also be applied to compare different survey samples [ 34 ].

In our analysis treatment modalities were defined as the two different questionnaires presented to the patient BSI versus OQ , and outcome was defined as the scale scores measured at intake and at discharge and then converted either in categories of clinical significance or in z values.

As a matching algorithm we applied the nearest neighbor matching with a logistic regression-based propensity score [ 35 ]. The following covariates were used in the regression equation: sex, age, marital status, educational level, GAF and CGI, respectively, at intake and at discharge, principal diagnosis, compulsory treatment order, duration of treatment, and type of discharge.

The adequacy of the matching procedure was checked using graphical visualizations of the propensity score distribution; the quality of the balance in the matched groups was examined through the standardized difference of means of each covariate [ 36 , 37 ]. Besides the comparison of the proportions of cases in the principal categories of clinical significance classified by the two instruments, we also analyzed the pre-post differences using the linear mixed model; for this, the OQ and BSI scores were z -transformed.

The completeness of the returned questionnaires varied between the two measures. The highest number of unanswered items was found for the OQ questionnaires returned by the inpatients. Questions on intimate relationships but also work-related questions were the most affected by nonresponse in sample 1. Overall, the impact of incomplete responses in sample 1 was high: of the questionnaires evaluable at intake only Nonresponse on the BSI questionnaires returned by inpatients was not related to the content but instead to the position of the items.

Items 1 to 10 had on average 1. Of the evaluable BSI forms at intake only In the outpatient sample the incompleteness of the questionnaires was low. This difference has to be viewed as improvement in data monitoring over the course of the years and cannot be considered to be an indicator for better acceptance of the BSI over the OQ The response rate was dependent on the severity of psychopathology at admission and of the amount of improvement at discharge in both inpatient samples Figure 2.

Relationship between missingness proportion of nonresponse and CGI ratings. The ratio between the number of respondents and the number of nonrespondents is noted on top of the bars data from samples 1 and 2. In the outpatient sample the response rate at the end of the therapy did not correlate significantly with Axis I or Axis II diagnoses. Table 2 shows the descriptive statistics of the scales based on data collected at the beginning of the treatment in the three samples.

Table 3 reports comparative statistics from German sample 3. Therefore, a patient can be classified as remitted if the following two conditions are fulfilled: a his OQ Total Score has decreased by at least 18 points and b has reached a value below 59 at the end of the treatment. If only b is met, then the patient is classified as improved.

Both parameter values are similar to those published in Puschner et al. Schmitz et al. Lutz et al. Figure 3 shows the profiles of different diagnosis groups on the OQ scales. The other two scales, that is, IR and SR, which have a minor importance in forming the Total Score, also did not vary substantially across the different type of disorders.

OQ profiles of the respondents at intake. For sample 3 only data of patients with a principal diagnosis of mood MD or anxiety disorder AD grouped according to the presence or absence of a personality disorder PD are represented. Percentages represent the sensitivity to psychopathology according to the OQ Total Score. BSI profiles of the respondents at intake. Percentages represent the sensitivity to psychopathology according to the GSI scale. In contrast to the OQ, the BSI produced profiles with more distinctive differences between the diagnosis groups Figure 4.

Except for the patients with schizophrenia F2 , the Depression scale was the scale with the highest score for the different diagnosis groups. The largest mean on this scale was exhibited by patients with a personality disorder F6 and not, as theoretically expected, by patients with an affective disorder F3.

Patients with schizophrenia attained the highest score on the Paranoid Ideation scale, which would be in accordance with the diagnostic criteria for this psychiatric group. However, their mean score was outperformed by that of patients with an F6 diagnosis. Overall, the highest Symptom Distress as measured by the BSI was observed on average among patients with a personality disorder in both an inpatient and an outpatient setting.

However, the same diagnosis groups with a low average OQ Total Score, that is, F2 and F1 among the inpatients and anxiety disorders without PD among the outpatients, achieved a GSI score that was on average also lower than that of other diagnosis groups.

Therefore, the substantial misclassification of patients with schizophrenia cannot be considered to be a consequence of low severity of mental illness or high psychosocial functioning of the sample analyzed. The statistics in Table 2 show that respondents with 0 points on one of the questionnaires were found among the inpatients but not among the outpatients and that the 5th and the 10th percentiles in the inpatient samples were lower than the corresponding values from the outpatients.

The lowest 5th percentiles of the global questionnaire scores were found for the GSI among the inpatients with substance abuse 0. For the inpatient setting, we based our analysis of sensitivity to change on matched samples. The histograms in Figure 5 show that samples 1 and 2 were already quite similar before the matching, since they exhibited an extensive overlap of their propensity score distributions but with some density differences, however.

From the smaller sample, that is, sample 1, a total of patients with complete scales and covariates values were available for the matching. Distributions of the propensity scores of the inpatient respondents with complete covariates values. Histograms on the left show the distributions before raw and after matched the matching. The plot on the right shows the differences between matched and unmatched cases. Figure 6 shows the results of the clinical significance analysis. In the outpatient sample, the proportions of improved and remitted cases classified by the two measures were quite similar.

Results of the analysis of the clinical significance. To balance out this difference, the scores of the two questionnaires were z -transformed and analyzed with a linear mixed model. Figure 7 shows the estimated fixed effects; they indicate that the two questionnaires recorded the same amount of change between the pre- and the postmeasurement.

Average pre-post changes in z scores. This study examined the applicability of the OQ and the BSI for assessing the outcome quality of inpatient and outpatient treatments. Both of these self-report measures can be easily administered by a wide range of service professionals and take about 10 minutes to be filled out.

Normative data and analysis results concerning their psychometric properties are available [ 38 ]. However, our analyses pointed out the following critical aspects of the performance of the two questionnaires, which have often been neglected in the literature: 1 the number of missing values that emerges in the data collection, 2 the diagnostic value of the scale profiles, and 3 the robustness of the clinical significance algorithm.

Since Rubin's [ 39 ] seminal paper on inference with missing data, there has been a growing awareness of this problem in the scientific community. In evaluation studies, the probability of nonresponse is often correlated with the attained outcome. Our results clearly demonstrate this relationship. Respondents and nonrespondents are different in two clinically crucial aspects: nonrespondents have higher severity of mental illness and show less improvement after the treatment than respondents.

Missingness is therefore a source of bias when assessing the effectiveness of a treatment, and nowadays guidelines concerning the statistical analysis of incomplete data are available [ 40 , 41 ].

In contrast, with samples 1 and 2 the application of multiple imputation proved to be ineffective. On the other hand, the drastic reduction of the data collection to only pre-post measurements, in order to minimize administrative expense, makes it hard to obtain robust estimations of the missing data through imputation models. All in all, the high proportion of missing data discourages the use of self-report measures with the patients with severe impairments usually found in a psychiatric hospital setting in favor of clinician-administered measures.

Questionnaires like OQ and BSI have a relatively large number of items, so that different reliable Likert scales may be formed that allow the creation of a person's profile. Can these profiles be used, for instance, to facilitate the formulation of a psychiatric diagnosis? Our results do not support the use of these questionnaires as screening instruments to facilitate the assessment of ICD diagnoses. In the construction of the OQ this was never an intended purpose [ 45 ], but the nine primary symptom dimensions of the BSI suggest a possible application for screening purposes.

In our sample, inpatients with personality disorders attained on average higher scores than inpatients in other diagnosis groups on six of the nine BSI scales. These patients had a higher mean score on the Depression scale than patients with an affective disorder and a higher mean score on the Paranoid Ideation scale than patients with schizophrenia.

Two different approaches have been suggested for dealing with the low discriminant validity of the BSI scales. The first is to consider the questionnaire through its GSI score as more appropriate for measuring the overall degree of psychological distress instead of the precise nature of the psychopathology [ 38 ].

From this optic, the Outcome Questionnaire with 45 items in its full version or 30 items in its short version would seem to be a more time-effective choice than BSI when measuring general level of psychological distress in a less time-consuming way.

The second approach consists in improving the factorial structure of the questionnaire. To this purpose, different authors have used the bifactor structural model in recent years. This model is used to build a general distress factor and more specific components of psychopathology. Thomas [ 47 ] demonstrated that a bifactor model of the BSI items can achieve higher accuracy than an oblique simple structure in diagnosing some disorders, such as depression or generalized anxiety disorder.

Brodbeck et al. One of their results in line with ours is that patients with personality disorders are characterized by a high level of general distress.

Overall, it seems that improving the factor structure of the BSI can lead to improved sensitivity in identifying depressive or anxiety disorders, but we doubt that it can do the same with patients with substance abuse or acute psychotic disorders.

A nonnegligible part of these patients tends to score low, and their profiles resemble those of healthy persons or remitted patients. These results support the hypothesis that patients with these disorders are inclined to underestimate their own emotional and behavioral difficulties.

As pointed out by Burlingame et al. Clinical significance, as originally proposed by Jacobson and Truax [ 19 ], is considered, beside pre-post effect sizes, to be a gold standard as a performance indicator in routine outcome monitoring [ 49 ].

This approach encompasses two steps: first, identifying the subsample of patients that reached a reliable change and then determining which among them moved outside the range of the dysfunctional population. This method presupposes a valid and consistent distinction between functionality and dysfunctionality already at the onset of the treatment. Therefore, the percentage of cases that can be adequately categorized depends on the sensitivity of the instrument.

Our results reveal a heterogeneous picture in this regard. Its reliability has been demonstrated in many Dutch and international studies. Thanks to these advantages, the test has become one of the most widely used measuring instruments in the world. The graph above shows the results by sub-scale at the start and completion of the trajectory.

This allows you to spot any underlying issues at a glance. The treatment effects of each dimension are expressed in T-score and Reliable Change.