Learning to live with missing quality-of-life data in advanced-stage disease trials.
Explore the source record for details and available documents.
Antineoplastic Combined Chemotherapy Protocols↗
Biomedical subjects
Gary W Donaldson
Publications and source records attributed to Gary W Donaldson.
Structural equation models for quality of life response shifts: promises and pitfalls.
Explore the source record for details and available documents.
Health Status↗
Correlation of change in visual analog scale with pain relief in the ED.
OBJECTIVES: To evaluate the validity of change in visual analog Scale (VAS) as a measure of pain relief using a verbal descriptor Scale (VDS) of change in pain. METHODS: A prospective observational study of emergency department patients measured pain with VAS and recorded verbal report of change in pain. RESULTS: One thousand four hundred ninety patients yielded 1999 comparisons between change in VAS and VDS. Correlation of change in VAS and VDS of change in pain was rho = 0.667 ( P < .001). A wide range of change in VAS, large standard deviations for the mean change in VAS, and discordance in the direction of change in VAS were present within each verbal descriptor category. CONCLUSIONS: Change in VAS is moderately correlated with a VDS of change in pain. Wide variability in change in VAS and discordance with a VDS demonstrate that change in VAS is not a valid indicator of pain relief for individual patients.
Adolescent↗
General linear contrasts on latent variable means: structural equation hypothesis tests for multivariate clinical trials.
Structural equation models articulate the assumed measurement and causal relations among variables, imposing discipline on otherwise unstructured and redundant associations that arise in correlational studies. Biomedical research has eschewed such methods, relying on the generally superior causal inference afforded by randomized controlled trials. Increasingly, however, clinical trials incorporate numerous covariates that are measured but unmanipulated. Most clinical trials now also include multiple correlated endpoints, which can generate ambiguous outcome patterns refractory to simple statistical analysis and interpretation. Modern clinical trials are really multivariate longitudinal studies with at best a component of randomized control; as such, structural equation approaches can add rigour and clarity. The analysis of latent variance (LANOVA) conception combines structural equation and experimental analysis of variance legacies. It allows, for any design that can be decomposed into between-group and within-person models, tests on latent means (that is, the means of the unobserved factors) that are directly analogous to their analysis of variance counterparts. LANOVA variables are either outcomes (which may have a time structure), varying covariates (which may have a time structure), or background covariates (which are static). Allowable causal relations are set-recursive: background covariates can affect all other variables; varying covariates can affect current outcomes and later outcomes and covariates; outcomes can affect only later outcomes. All standard ANOVA and ANCOVA hypotheses can then be tested by proper restrictions among the means and intercepts of these latent variables. Structural equation modelling programs can be used to estimate the models and test the hypotheses. I demonstrate the approach by expressing and testing hypotheses appropriate for two clinical studies that evaluate patient-reported outcomes in populations with solid and haematological malignancies.
Analysis of Variance↗
Pain and the defense response: structural equation modeling reveals a coordinated psychophysiological response to increasing painful stimulation.
The defense response theory implies that individuals should respond to increasing levels of painful stimulation with correlated increases in affectively mediated psychophysiological responses. This paper employs structural equation modeling to infer the latent processes responsible for correlated growth in the pain report, evoked potential amplitudes, pupil dilation, and skin conductance of 92 normal volunteers who experienced 144 trials of three levels of increasingly painful electrical stimulation. The analysis assumed a two-level model of latent growth as a function of stimulus level. The first level of analysis formulated a nonlinear growth model for each response measure, and allowed intercorrelations among the parameters of these models across individuals. The second level of analysis posited latent process factors to account for these intercorrelations. The best-fitting parsimonious model suggests that two latent processes account for the correlations. One of these latent factors, the activation threshold, determines the initial threshold response, while the other, the response gradient, indicates the magnitude of the coherent increase in response with stimulus level. Collectively, these two second-order factors define the defense response, a broad construct comprising both subjective pain evaluation and physiological mechanisms.
Adolescent↗
Pain medication use before ED arrival.
The objective of this study was to determine the frequency and types of pain medications taken before ED arrival based on pain intensity, duration of pain, chief complaint, gender, age, and race. A convenience sample of patients in pain was enrolled in this university hospital-based prospective, observational study. A total of 1233 patients were enrolled. Five hundred thirty-nine of 1233 (44%) patients took pain medication before arrival. Two hundred three (38%) took ibuprofen, 147 of 539 (27%) took oral opioids, and 135 of 539 (25%) took acetaminophen, which were the most frequently used medications. Severity of pain, age, duration of pain, and chief complaint were associated (chi-squared P <.05) with variations in prior medication use. Race and gender were not associated (chi-squared P >.05) with differences in medication use before arrival. Many patients (44%) take medication before arrival in the ED. Age, severity and duration of pain, as well as chief complaint are associated with differences in frequency of self-administered medication.
Age Factors↗
A psychophysiological causal model of pain report validity.
The validity of the pain report is vitally important but difficult to assess because pain is a personal experience. Human laboratory research affords an opportunity to investigate validity because one can measure the consistency and sensitivity of pain ratings produced in response to known stimuli. This article presents 2 levels of evidence characterizing the validity of the pain report measure. The within-subject agreement of pain report with known stimulus variation quantifies the criterion validity, or accuracy, of the measure. Causal modeling defines a second, between-subject, level of construct validity by suggesting a psychophysiological mechanism determining the observed individual variation in accuracy. We analyzed pain rating data obtained in a laboratory study where 100 subjects (56 men and 44 women) experienced varied levels of painful fingertip electrical stimulation, delivered in random order across 144 trials. Unknown to the subjects, there were only 3 stimulus intensities. Accuracy, defined operationally as the proportion of variance in pain report explained by stimulus level, ranged from 0.07 to 0.91 with a median of 0.64. Hypothesized determinants of accuracy comprised current intensity, event-related late near field evoked potentials, skin conductance response, heart rate, and pupil diameter change. We limited the evoked potential measures to the amplitude of the negative peak at 150 msec (N150amp) and combined the latter 3 measures to form a single index of overall sympathetic nervous system arousal (Arousal). Although men chose higher stimulus levels for the experiment and had higher Arousal than did women, their mean pain reports and their Accuracy did not differ from those of female subjects. We constructed a sequence of path analysis models designed to clarify the causal contributions of current intensity, N150amp, and Arousal, and to determine whether these relationships differ in men and women. The final model revealed a direct causal chain. Stimulus current determined the amplitude of N150amp (possibly an indicator of attention). N150amp in turn determined Arousal, and Arousal emerged as the sole determinant of the Accuracy of the pain report. In addition, this latter effect differed across the sexes. Men who experienced higher levels of Arousal gave more accurate pain reports than those who had lower levels, but women who had higher levels of Arousal gave less accurate pain reports than those with lower levels. Thus construct validation emerged, not from direct stimulus-response correlation, but from the elucidation of a causal chain that related stimulus to response.
Journal Article↗
Individual differences in quality-of-life treatment response.
BACKGROUND: Individual differences in treatment responses to health-related quality-of-life interventions arise ubiquitously and prominently in clinical trials. These differences do not reflect error, but instead represent patterns of response that vary reliably across patients. Individual differences complement and qualify the information conveyed by average health-related quality-of-life effects for treatment arms. If clinicians and patients are to use health-related quality-of-life findings from clinical trials to make treatment decisions, they must have information about the extent and nature of such individual variation. The relatively small effect sizes of health-related quality-of-life outcomes in clinical trials offer a weak basis for generalizing results to new persons, and patients routinely anticipate likely treatment benefits for themselves that exceed these small effects. A focus on individual differences in treatment response can promote more realistic appreciation of expected benefit and uncertainty. A pharmacogenetic example shows how individual differences in drug metabolism can directly affect health-related quality-of-life treatment outcomes such as pain and physical functioning. MEASURES: The authors suggest how graphical displays can summarize individual responses and provide a context for interpreting the size and generality of the average treatment. Mixed-effects modeling subsumes average treatment differences and individual differences in a unified statistical analysis. Analysis of health-related quality-of-life data from an advanced colorectal cancer trial illustrates this approach. Objective statistical criteria indicate that, for this example, individual differences dominate the treatment difference. CONCLUSIONS: The authors suggest that presentation of the spectrum of individual responses and associated prediction intervals can convey clinically meaningful information regarding the impact of a treatment on health-related quality of life.
Antineoplastic Agents↗
The factorial structure and stability of the McGill Pain Questionnaire in patients experiencing oral mucositis following bone marrow transplantation.
The McGill Pain Questionnaire (MPQ) (Melzack 1975) is an important assessment tool for multidimensional pain measurement in both clinical practice and research. Despite widespread acceptance, empirical analyses have not consistently verified the 3 a-priori factors that guided the subclass construction of the Pain Rating Index (PRI) of the MPQ. This study compared the a-priori model with 2 qualitatively different factor models in 191 patients with oral mucositis pain at 3 days and 10 days following bone marrow transplantation. A semantic model defined by Sensory Action, Sensory Evaluation, and Affective Evaluation factors of subclass descriptor content fit better than the a-priori model and a model positing a single general pain factor. The 3 semantic PRI factors were highly intercorrelated, with the sensory factors correlating more highly with an independent visual analogue (VAS) pain scale. Standardized factor regression coefficients between the two occasions of measurement ranged between 0.4 and 0.5. Mean factor change was greatest for Sensory Evaluation and lowest for Affective Evaluation. All analyses were conducted with the LISREL 7 structural equation modeling program. Although the factor analyses indicated an unambiguous ranking of PRI models according to statistical criteria, these theoretical results generalize poorly to simple scores formed by direct addition of the PRI subclasses. Summary scores can only approximate the unobserved factors and cannot retain the fine discriminations revealed by the theoretical factors. Psychometric considerations suggest that a single PRI total score will yield better practical measurement than any scoring rules based on multiple factors.
Adolescent↗
Relaxation and imagery and cognitive-behavioral training reduce pain during cancer treatment: a controlled clinical trial.
Few controlled clinical trials of psychological interventions for cancer pain relief exist in spite of frequent support for their importance as adjuncts to medical treatment. This study compared oral mucositis pain levels in 4 groups of cancer patients receiving bone marrow transplants (BMT): (1) treatment as usual control, (2) therapist support, (3) relaxation and imagery training, and (4) training in a package of cognitive-behavioral coping skills which included relaxation and imagery. A total of 94 patients completed the study which involved two training sessions prior to treatment and twice a week 'booster' sessions during the first 5 weeks of treatment. Results confirmed our hypothesis that patients who received either relaxation and imagery alone or patients who received the package of cognitive-behavioral coping skills would report less pain than patients in the other 2 groups. The hypothesis that the cognitive-behavioral skills package would have an additive effect beyond relaxation and imagery alone was not confirmed. Average visual analogue scale (VAS) report of pain within the therapist support group was not significantly lower than the control group (P = 0.103) nor significantly higher than the training groups. Patient reports of relative helpfulness of the interventions for managing pain and nausea matched the results of VAS reports. From these results, we conclude that relaxation and imagery training reduces cancer treatment-related pain; adding cognitive-behavioral skills to the relaxation with imagery does not, on average, further improve pain relief.
Adaptation, Psychological↗
Hypnosis or cognitive behavioral training for the reduction of pain and nausea during cancer treatment: a controlled clinical trial.
Few controlled clinical trials have tested the efficacy of psychological techniques for reducing cancer pain or post-chemotherapy nausea and emesis. In this study, 67 bone marrow transplant patients with hematological malignancies were randomly assigned to one of four groups prior to beginning transplantation conditioning: (1) hypnosis training (HYP); (2) cognitive behavioral coping skills training (CB); (3) therapist contact control (TC); or (4) treatment as usual (TAU; no treatment control). Patients completed measures of physical functioning (Sickness Impact Profile; SIP) and psychological functioning (Brief Symptom Inventory; BSI), which were used as covariates in the analyses. Biodemographic variables included gender, age and a risk variable based on diagnosis and number of remissions or relapses. Patients in the HYP, CB and TC groups met with a clinical psychologist for two pre-transplant training sessions and ten in-hospital "booster" sessions during the course of transplantation. Forty-five patients completed the study and provided all covariate data, and 80% of the time series outcome data. Analyses of the principal study variables indicated that hypnosis was effective in reducing reported oral pain for patients undergoing marrow transplantation. Risk, SIP, and BSI pre-transplant were found to be effective predictors of inpatient physical symptoms. Nausea, emesis and opioid use did not differ significantly between the treatment groups. The cognitive behavioral intervention, as applied in this study, was not effective in reducing the symptoms measured.
Adult↗
Strengthened estimates of individual pain trends in children following bone marrow transplantation.
Pre-transplant conditioning regimens for bone marrow transplantation often cause oral mucositis and severe pain. We evaluated the agreement of self- and parent reports of daily oral mucositis pain in children between the ages of 6 and 16 years. Child patients were asked to report their pain on visual analog scales (VAS) daily for 20 days following their transplants. Daily VAS ratings were also obtained from one of the parents. The analysis sample consisted of ten children aged 7-9, nine children aged 10-12, eight children aged 13-16, and their parents. We modeled individual child and parent reports as quadratic functions of the number of days post transplant. Empirical Bayes/restricted maximum-likelihood estimates were obtained of individual coefficients, treated as random effects, and age group coefficients, treated as fixed effects. Parents exhibited higher average pain curves than their children in each of the three age groups. The middle age group reported the highest average pain. Average within-person error variances representing unreliability were 692.2, 461.9, and 303.9 for young, middle, and old children, respectively; for parents, the corresponding error variances were 375.1, 413.3, and 252.4. These results challenge the presumption that children tend to over-report pain but are consistent with the contention that younger children may be less reliable reporters than adolescents and adults.
Adolescent↗