|
Advertisement |
|
|
||
 |
|
|||||
|
|||||
|
 |
||||||
|
||||||
|
||||||
|
© 2003 American Society for Clinical Oncology Feasibility and Compliance of Automated Measurement of Quality of Life in Oncology Practice
From Cancer Research UK, Clinical Centre in Leeds, St James’s University Hospital, University of Leeds, Leeds; Department of Medical Oncology, Airedale General Hospital, Keighley, England; Cancer Research UK Medical Oncology Unit, Western General Hospital; and Information and Statistics Division, National Health Services Scotland, Edinburgh, Scotland, United Kingdom. Address reprint requests to E.P. Wright, PhD, Cancer Research UK, Clinical Centre in Leeds, St James’s University Hospital, Beckett St, Leeds, England, LS9 7TF; email: medepw{at}cancermed.leeds.ac.uk.
Purpose: Systematic quality-of-life (QOL) assessment may have value in oncology practice by increasing awareness of a wide range of issues, possibly increasing detection of psychologic morbidity, social problems, and changes in physical status, and improving care and its outcomes. However, logistic problems are substantial. Automated systems solve many of these problems. We field-tested the feasibility and compliance that can be achieved using a computer touchscreen system in two consecutive studies. Patients and Methods: In study 1, a prospective cohort of 272 patients was offered QOL assessment at each clinic appointment for 6 months. In study 2, all patients (N = 1,291) were offered QOL assessment as part of clinic routine during a 12-week period. Results: In study 1, 82% of patients agreed to take part, but over time, compliance was poor (median, 40%; mean, 43%) and deteriorated with longer follow-up. In study 2, the overall compliance was greatly increased (median, 100%; mean, 70%), and compliance was retained over multiple visits. In study 1, compliance was better in younger patients, males, and socially advantaged patients, but was not affected by the presence of depression or anxiety, or QOL. In the second study, building on experience in the first study, data collection and storage in the computer system was excellent, achieving 98% of collected data stored in one center. In general, patients were comfortable with the computers and the approach. Data collection on the wards was more difficult and less complete than in clinics, especially for patients undergoing acute admissions. Conclusion: Feasibility with higher compliance was demonstrated in study 2, in which the data collection was integrated into routine care, and can be improved with further technical initiatives and education of staff.
QUALITY OF LIFE (QOL) is now established in cancer clinical trials as a principal or subsidiary end point, and we are beginning to appreciate the clinical meaning and value of these data.1–4 However, the value of the direct measurement of QOL in oncology practice also may be substantial.5–9 It may enhance the interaction between healthcare workers and cancer patients by increasing awareness of a wide range of issues, including detection of psychologic morbidity and assessment of functional status and physical symptoms, and therefore, it can influence the process of care and its outcomes. The discovery of problems and issues may be increased by the use of QOL instruments and the presentation of those data to healthcare workers. Low detection rates for anxiety, depression, and other psychologic distress and inappropriate referral to support services are recognized in oncology.10–12 Self-report screening questionnaires for psychologic distress have been tested with cancer patients, and the detection rates, although not high, are better than those found in standard clinical practice.13 The introduction of routine screening for psychologic distress should improve detection and may result in more appropriate referrals and improved outcomes, although this has not yet been demonstrated.14 There remain barriers to overcome before QOL measures and psychosocial screening can be established routinely in oncology and lead to benefits for patients. The manual collection of these data, their analysis, entry into databases, and presentation on paper are difficult, laborious, and consume substantial time and resources. It remains uncertain whether patients will be willing to comply with routine QOL data collection. Typically, approximately 70% of the patients in clinical trials complete the baseline QOL measurement, but with follow-up assessments, the compliance is lower and missing data are a problem. Compliance with QOL assessment in clinical trials with regular support from research staff may differ from compliance with routine QOL data collection in practice, in a busy oncology clinic. Unlike other clinic procedures (weighing, blood tests, and so on), the noncompletion of the QOL assessment might not be expected by patients to hinder medical consultation or prevent treatment, so it might be perceived as unnecessary. We feel that for the field to progress, the logistic problems must be overcome and feasibility and compliance assessed. The collection of QOL data should become robust, inexpensive, easy, and readily interpretable. Its place in clinical practice can then be evaluated in large patient numbers and different clinical settings. For this reason, we and others have developed and evaluated automated systems.15–19 The ideal system will allow direct input from patients to measure their responses to standardized questionnaires of appropriate reliability, validity, and scope of inquiry. The data should automatically be entered into a database, analyzed, and presented to healthcare workers in real time to make it possible that the process of care is immediately informed. Although this ideal has not yet been achieved, we have shown that the use of computer touchscreens with relatively brief cancer-specific questionnaires (the European Organization for Research and Treatment of Cancer Quality of Life Questionnaire-Core 30, version 3.0 [EORTC QLQ-C30]2 and the Hospital Anxiety and Depression Scale [HADS]20) allows the collection of data quickly and easily, and that these data are usually the same as those collected by paper and pencil approaches.17 This approach is acceptable to patients and will discover additional information to that found in conventional medical notes. There is evidence that automated QOL data collection will alter the process of care, although there are not yet sufficient data for us to conclude that it can improve outcomes for our patients.9,21 The broad aims of the projects were to introduce automated regular measurement of QOL in oncology clinical practice and to examine the feasibility of the approach in terms of patient willingness to participate in the project, compliance over time, and attitude to the measurement. The use of QOL results by the professionals was not the focus of this study, but some information was collected and limited results are reported here. In this article, we report our experience of field-testing the feasibility and compliance that can be achieved with computer touchscreen systems and cancer patients in clinical practice. First, we assessed compliance in a cohort of patients who gave informed consent to enter the study, and the data that related to their longitudinal follow-up was collected (study 1). This study ran from April 1997 to June 1998. The results of this study next led us to try a strategy in which all patients in a practice (clinic and ward based) were asked to complete the touchscreen systems at each visit, and the data that related to their longitudinal follow-up were extracted from the total data pool—an all-comers approach (study 2), undertaken in 1999. For this latter approach, we were given permission by our institutional ethics committees to offer patients the QOL assessments without prior individual informed consent. Study 1 was performed in two centers, St James’s Hospital, Leeds, and Western General Hospital, Edinburgh, United Kingdom, and, in addition, at Airedale Hospital, West Yorkshire, United Kingdom, in study 2.
Studies 1 and 2 In both studies, the instruments were the EORTC QLQ-C30, which comprises 30 questions, including functioning scales (physical, role, emotional, cognitive, and social) and symptom scales and items, and the HADS, a 14-item questionnaire designed specifically for detecting anxiety and depression in physically ill people. The questionnaires were presented to patients via a desktop computer touchscreen. For security, the hardware was locked to a trolley in Edinburgh and housed in locked cabinets in Leeds and Airedale. The computers were used by patients in quieter areas of the clinic to try to ensure privacy. A computer program was written using Visual Basic 5 (Microsoft Corp., Seattle, WA) for presenting the questionnaires on the computer touchscreen. The program included storage of the raw scores on an MS Access 2 (Microsoft Corp., Seattle, WA) database and conversion of the scores to scale score for paper printout. Questions were presented on the touchscreen one question at a time using the same instructions as with the paper versions. A bar code reader allowed patients to access the program using bar codes derived from the patient’s name and hospital number. Once the patients had completed the questionnaires, the computer program scored the results automatically, and an immediate printout of aggregated scores was available. Patients were told the purpose of completing the questionnaires was to provide information about the feasibility of introducing routine QOL assessment and that this would contribute toward planning patient-centered care in clinics. No assurance was given to patients about staff using the QOL results, and patients were encouraged to report any concerns or difficulties to the healthcare team as usual.
Study 1: Prospective Cohort Study In both centers, patients were recruited and followed up for 6 months. At the time of consent, patients were given a bar code that was attached to their appointment card. The patients were instructed how to use the bar code reader and how to complete the questionnaires using the touchscreen, which took up to 5 minutes per patient. This consent-baseline visit was supervised by research staff, and further supervision was not provided at follow-up visits unless requested by the patient. Patients were asked to complete the two questionnaires via the touchscreen every time they attended the outpatient oncology clinic during 6 months unless they visited more frequently than once a week. An immediate printout of the results from each visit was available. However, doctors were free to choose whether to use the QOL results. Later, the results were attached to a QOL mount sheet, designed specifically for QOL reports in the investigations section of the medical notes, by the clinic nursing or clerical staff, according to usual procedures. At visits after the consent visit, the clinic patients were not prompted by the staff to complete the questionnaires; the responsibility was placed entirely on the patient, and no feedback was given concerning use of the QOL results.
Study 2: All-Comers Study Procedures differed somewhat in the participating centers. In Leeds, the system was networked through the hospital network to a printer. In Edinburgh, printing was local to the clinic, and the printout was attached without comment to the case notes before they were taken into the consultation. In Leeds and Airedale, clinic and ward staff were asked to collect the printout in the same way as other results and attach it to the front of the medical notes, and these were to be signed by the medical oncologists as is usual procedure for all test results before filing. An information sheet was given to each clinic member of staff before the study began. Medical professionals attending to the patient had the printout of QOL scores available to them. At the end of each patient consultation, doctors in Leeds and Airedale were asked to complete a short questionnaire on the usefulness of the QOL results for the respective patient. Physicians were asked to indicate whether they used the QOL data, whether they found it clinically useful, and, if yes, in what way (providing information, confirming knowledge, identifying problems, contributing to management, and so on).
Data Analysis Logistic regression analyses were fitted using generalized estimating equations,23 which related the probability of compliance on each visit to the baseline demographic and psychologic measurements taken on each patient. Sex, marital status, Carstairs index, education, and response to treatment at 3 months were fitted as classification factors, and the other variables were fitted as mean-centered continuous covariates. A term was included for the visit number, with the first visit coded as 0. Where significant differences were found, the continuous covariates were categorized to explore the nature of the differences. The generalized estimating equations analysis extends fixed effects logistic regression analysis24 by estimating a term for the correlation between the compliance on each visit for the same patient, assuming independence of compliance among different patients. The within-patient correlation was assumed to take a first-order autoregressive form, which implies that the results for sequential visits are more highly correlated than for visits further apart. Study 2. In study 2, at the end of the 3 months, all patient appointments or admissions were counted. The number of QOL assessments completed was recorded. The total number of results filed in the medical notes was checked (Edinburgh excluded). The number and dates of completed questionnaires were recorded from the QOL database. Global participation was calculated as the number of patients completing QOL assessment divided by the total number of patients in each clinic; compliance at each visit had to be derived from the total collected data rather than a selected prospective cohort. Compliance was calculated for each visit by dividing the number of times that a QOL assessment was completed by the number of possible completions at that visit; for example, for visit 3, the denominator was the total number of patients recorded as having attended a third time, and the numerator was the number of QOL assessments completed at a third visit. Descriptive statistics were used to analyze doctors’ views on the clinical usefulness of QOL data for their individual patients.
Studies 1 and 2.
Patients were asked about the equipment, the clinic routine, the usefulness of the questionnaires, and their overall attitude toward routine QOL assessment. The number of QOL printouts stored in the medical notes and the number of complete data entries on the computer touchscreen were counted and compared, and each center gave a description of any problems encountered with the system. Patient groups (participants, refusers, and centers) were compared in terms of demographic and clinical variables using the
Study 1 Compliance at Baseline and Over Time During the study 1 period, 6% of 345 patients who attended the outpatient clinics were not approached because of ineligibility. Among the 324 patients approached, 272 consented and 52 refused. A total of 1,250 patient appointments were attended over the study period by the consenting group, ranging from some patients attending only once to one patient attending 19 times, including the consenting visit. Seven consented and then did not complete an assessment; therefore, 265 patients consented and completed an assessment. A baseline compliance estimate (assessed/approached) was thus 82% (265 assessed of 324 approached) in study 1. The patients who consented were compared with those who refused by age, sex, and disease progression, and no significant differences were found between the two groups. Reasons for refusal were felt too ill (n = 10), could not be bothered (n = 22), felt too distressed (n = 7), did not like computers (n = 4), and other (n = 9).
The majority of patients partially complied; that is, they completed the QOL assessment at some appointments but not at others. There was a large group who did not complete any QOL assessments after the baseline visit and a smaller group who completed the QOL assessment at every visit (Fig 1A
Study 2 Compliance at Baseline and Over Time During the 3 months of study 2, 1,291 patients attended the clinics, some of whom attended more than once, resulting in a total number of patient visits of 1,826. Patients are described in Table 1  . QOL assessments were completed on 1,271 of the 1,826 patient visits and not completed on 555 patient visits. Among these 555 noncompletions, 333 patients refused, of whom 210 gave a simple refusal ("do not want to" or "can’t be bothered"), 78 were too unwell physically or emotionally, eight were in another QOL trial, seven disliked computers, and 30 stated various other reasons. Forty were felt to be ineligible (most common reasons were "too ill" or unable to speak English). In 100 patients, data were not collected because other aspects of the clinic visit took precedence (called to see staff or have tests). Computers malfunctioned in 22 cases. Global compliance was better in men (P = .043) and younger people (80%, younger than 40 years; 71%, 40 to 60 years; and 64%, older than 60 years).
A substantial proportion of patients complied completely, although unlike in study 1, these included a large number of patients who attended on one occasion only. There were a few patients who partially complied and a significant minority who did not complete the touchscreen questionnaires at all (Fig 1B  ). The overall compliance was greatly increased in comparison to study 1, (mean, 72%; median, 100%) and the distribution was altered.
Fig 1A For study 2, overall compliance at two Leeds clinics, one clinic in Airedale Hospital, and the Edinburgh clinics was compared. The difference in clinic compliance was significant (P < .001), with the nurse-led adjuvant chemotherapy clinic at Leeds having the highest compliance, with a mean of 93%, and the other clinics having a mean compliance of 65% to 70%.
Study 1: Factors Predicting Compliance
Compliance was better in younger patients, male patients, and patients with less deprivation, as measured by the Carstairs index. There was some evidence to suggest that married and divorced, separated, or widowed patients were more compliant than single persons. Depression, anxiety, and QOL scores were not associated with differences in compliance. None of the functional scales were significantly statistically related to compliance. Education and response to treatment at 3 months were not significantly statistically related to compliance.
Studies 1 and 2 Follow-Up and Compliance Over Time
For this comparison, we have corrected the patient numbers for the ineligible patients identified in study 2. Thus, 1,291 patients (all in the study) made one visit and 27 were ineligible; therefore, the denominator for the first visit is 1,264 patients. The inclusion or exclusion of eligible patients does not alter the conclusions.
Study 1 Patient Feedback
Study 2 Patient Feedback
Table 3
Study 2 Doctors’ Views on the Clinical Usefulness of the QOL Data The data were collected on an exploratory subsample only for the patients in Leeds and Airedale. Thirteen physicians participated. They completed 315 questionnaires (86%) out of 366 individual consultations during the study period. In 41% of the visits, the doctors used the information (responses "somewhat, quite a bit, or very much"), and in 26% or the visits, they used it only "a little." Similarly, in 42% of the visits, physicians found the QOL data clinically useful ("somewhat, quite a bit, or very much"), and in 21% of the visits, they found it "a little" useful. The QOL data provided additional information in 24% of the visits, confirmed doctors’ knowledge of patient problems in 24% of visits, identified issues for discussion in 14% of visits, and contributed to patient management in only 5% of the consultations.
Study 1 Data Collection Problems
Study 2 Data Collection Problems
Study 2: Ward Study Fifty-six percent of ward patients returned their feedback questionnaires. Although the majority did not have any previous experience with computers, none reported significant problems with the computer touchscreen. As with outpatients, an overwhelming majority felt that the questionnaires and touchscreens had not interfered with their hospital stay and had not caused them any worry or distress.
The following conclusions can be drawn from the two studies: (1) It is feasible to use the computer touchscreen systems to generate data on a high proportion of large numbers of patients attending oncology clinics and to a lesser degree on wards. (2) An approach that incorporates data collection into routine clinical practice generates better compliance overall and, in particular, better data to describe changes over time than an approach that identifies a separate cohort prospectively. (3) Patients, in general, find the equipment easy to use with few problems, and their feedback is generally positive. (4) Experience and development of hardware and software improves results. Although there is room for further improvement, current completeness of data (study 2) is good. Because of the differences in study design, direct comparison of compliance between the two studies is difficult. Study 2 represents our best attempt to resemble, as closely as possible, clinic practice; the result that compliance did not decrease below 60% is encouraging. Study 2 also showed robust data collection and data storage systems drawing on the experience gained in the earlier study. The main reason for noncompliance was simple refusal. More than half of the refusers stated they could not be bothered to complete the assessment, with others feeling too ill or too distressed. Although it was encouraging that depression, anxiety, functional scales, and overall QOL did not predict for poor compliance, compliance was not randomly distributed; it was better in younger and economically advantaged patients. The reasons for refusal and the relation of social deprivation and age to noncompliance are of some concern. It may be that some people with a high level of need would not benefit from introduction of routine QOL assessment because of nonparticipation. It is difficult to determine how further developments in the system will address this problem, but a reason for optimism is found in the literature. In a review of studies that examined the adoption of new innovations in organizations, Rogers25 found that late adopters (or, in our studies, the noncompliers) were more likely to have less education and have lower social status than early adopters. However, the adoption of innovation usually occurs even with the late-adopter group over time. If routine assessment of QOL becomes commonplace in oncology outpatient clinics, some late adopters may gradually accept routine QOL assessment. This is an area that requires further scrutiny and will continue to be assessed in future studies. These automated and systematic methods for improving the care of cancer patients will complement, not replace, standard patient assessment. Administrative error accounted for 28.6% of nonparticipation, with most patients being missed by the researcher during the clinic visit. This should improve with experience and the organization of this approach in clinics and wards. If only eligible patient visits are examined, the compliance rate increased to a mean of 71%. The nurse-led adjuvant clinic outperformed the other places of data collection, possibly reflecting the way the clinic is run. There are relatively low numbers of patients attending, patients attending tend to be younger and less ill, and patients are in a phase of active therapy. Establishing the QOL assessment on the ward was more difficult than at the clinics. The global rate of compliance was much lower than at the clinics (43.1%). There seemed to be two main reasons for this: (1) 37% of ward admissions were acute—of these, 73% were ineligible for the study; and (2) because of varied admission times and demands on patient time at admission, it was much harder for the research team to catch the patient at the most opportune moment, resulting in 10% of patients not participating. There was a highly significant difference between the levels of compliance on elective admissions and acute admissions. If we remove the noneligible patient visits, compliance increases to a mean of 67.9%. This compliance rate is steady over ward visits one and two (where numbers were greater than 12).
The change in the distribution of compliance among patients between study 1 (Fig 1A Although the context of these studies is very different from that of a randomized prospective trial, it is interesting to compare the compliance reported in randomized trials. This issue was discussed extensively in a 1998 issue of Statistics in Medicine.26–36 Compliance depends on study design and the approach to data collection, follow-up, and repeated questioning. The best results for compliance over time in randomized prospective trials maintained the completeness of follow-up at more than 80% for some years.26 However, commonly in the routine clinical trials context compliance will decrease below 70% after a year. It is difficult to make a comparison to the present studies, particularly to study 2, where we evaluated an unselected population of patients attending our clinics and there was no prospective informed consent. In the setting of a clinical trial, willingness to enter the study and to complete the QOL questionnaires will be confirmed at the beginning when consent is given. In our clinical practice setting, the group of refusers were counted but would be excluded from a trial by refusing to give informed consent. Feasibility and compliance are good enough to persevere with the goal of using automated QOL data collection to address some of the clinical challenges we have discussed. In particular, the results in study 2 are sufficiently good to be the basis of a successful program of work in this area. It is likely that compliance, and therefore the utility of these approaches, will improve as the all-comers approach is steadily built into the routine of clinics and, with greater difficulty, wards. This will rely on clinical or nursing staff taking on the task of introducing the assessment to patients, which may influence compliance initially until the system becomes fully integrated into practice. Our feedback suggests that there is no lack of willingness to include QOL assessment in routine practice among staff or patients. Over time, staff members are likely to become more familiar with interpreting QOL results and possibly will make more use of them. This may lead to patients (more than half of whom believed that the QOL results did not contribute to their consultation in study 2) beginning to see changes in the patient-doctor consultation. This should lead to benefits for both staff and patients, possibly leading to increased compliance. Developments in organization and technology can only be one part of the picture. Formal evaluation is necessary to determine whether patients will benefit from the availability of these additional data to their healthcare professionals. Our preliminary data9,21 and work in lung cancer suggest that there will be improvements in the process of care,18 but final demonstration of improved outcomes must await the results of large randomized prospective trials. We have completed the accrual of patients into such a trial to evaluate the improvements in healthcare that may come from repeatedly measuring QOL in practice and returning these data to healthcare professionals. The impact of QOL data depends on the commitment, training, and skills of the healthcare professionals to whom the data are presented. This remains a new area, and benefits for patients should increase as experience and training for healthcare cancer professionals in the use of QOL and related data increases. We believe that the collection of QOL data may enhance the interaction between healthcare professionals and patients, usefully screen for psychologic distress and social problems, and contribute to assessment of functional status and physical symptoms. Work is necessary to develop appropriate approaches in each of these areas and test their clinical value.
Carstairs scores were provided by the Office of National Statistics, London, United Kingdom.
Supported by a grant from the National Health Services National Research and Development Programme (E.P.W., A.S., and K.W.), Cancer Research UK (G.V., A.C., and P.J.S.), and the National Lotteries Charities Board (G.V.).
1. Selby P, Chapman JW, Etazadi-Amoli J, et al: The development of a method for assessing the quality of life of cancer patients. Br J Cancer 50:13–22, 1984[Medline]
2. Aaronson NK, Ahmedzai S, Bergamn B, et al: The European Organization for Research and Treatment of Cancer QLQ-C30: A quality of life instrument for use in international clinical trials in oncology. J Natl Cancer Inst 85:365–376, 1993 3. Osoba D: Lessons learned from measuring health-related quality of life in oncology. J Clin Oncol 12:608–616, 1994[Abstract]
4. Osoba D, Rodrigues G, Myles J, et al: Interpreting the significance of changes in health related quality of life scores. J Clin Oncol 16:139–144, 1998 5. Ganz P: Quality of life and the patient with cancer individual and policy implications. Cancer 74:1445–1452, 1994[CrossRef][Medline] 6. Detmar SB, Aaronson NK: Quality of life assessment in daily clinical oncology practice: A feasibility study. Eur J Cancer 34:1181–1186, 1998[CrossRef][Medline] 7. Lohr KN: Applications of health status assessment measures in clinical practice: Overview of the third conference on advances in health status assessment. Med Care 30:MS1–MS14, 1992 (suppl)[CrossRef][Medline]
8. Higginson IJ, Carr AJ: Using quality of life measures in the clinical setting. BMJ 322:1297–1300, 2001 9. Velikova G, Brown J, Smith A, et al: Computer based quality of life questionnaires may contribute to doctor patient interactions. Br J Cancer 86:51–59, 2002[CrossRef][Medline] 10. Cull A, Stewart M, Altman DG: Assessment of and intervention for psychosocial problems in routine oncology practice. Br J Cancer 72:229–235, 1995[Medline] 11. Ford S, Fallowfield L, Lewis S: Can oncologists detect distress in their out-patients and how satisfied are they with their performance during bad news consultations? Br J Cancer 70:767–770, 1994[Medline] 12. Ramirez AJ: Liaison psychiatry in a breast cancer unit. J R Soc Med 82:15–17, 1989[Abstract] 13. Ibbotson T, Maguire P, Selby P, et al: Screening for anxiety and depression in cancer patients: The effects of disease and treatment. Eur J Cancer 30A:37–40, 1994[CrossRef][Medline]
14. Gilbody S, House A, Sheldon T: Routinely administered questionnaires for depression and anxiety: Systematic review. BMJ 322:406–409, 2001 15. Sigle J, Porzsolt F: Practical aspects of quality of life measurement: Design and feasibility study of the quality of life recorded and the standardised measurement of quality of life in an outpatient clinic. Cancer Treat Rev 22:75–89, 1996 (suppl A)[Medline] 16. Buxton J, White M, Osoba D: Patients’ experiences using a computerized program with a touch-sensitive video monitor for the assessment of health related quality of life. Qual Life Res 7:513–519, 1998[CrossRef][Medline]
17. Velikova G, Wright EP, Smith AB, et al: Automated collection of quality-of-life data: A comparison of paper and computer touch-screen questionnaires. J Clin Oncol 17:998–1007, 1999 18. Taenzer P, Bultz BD, Carlson LE, et al: Impact of computerized quality of life screening on physician behaviour and patient satisfaction in lung cancer outpatients. Psychooncology 9:203–213, 2000[CrossRef][Medline] 19. Cull A, Gould A, House A, et al: Validating automated screening for psychological distress by means of computer touchscreens for use in routine oncology practice. Br J Cancer 85:1842–1849, 2001[CrossRef][Medline] 20. Zigmond AS, Snaith RP: The Hospital Anxiety and Depression Scale. Acta Psychiatr Scand 67:361–370, 1983[Medline]
21. Velikova G, Wright EP, Smith A, et al: Self-reported quality of life (QL) of individual cancer patients: Concordance of results with disease course and medical records. J Clin Oncol 19:2064–2073, 2001 22. Carstairs V, Morris R: Deprivation and Health in Scotland. Aberdeen, United Kigdom, Aberdeen University Press, 1991
23. Liang KY, Zeger SL: Longitudinal data analysis using generalized linear models. Biometrika 73:13–22, 1986 24. Nelder JA, Wedderburn RWM: Generalized linear models. J R Stat Soc A 135:370–384, 1972 25. Rogers EM: Innovations and Adopter Categories, in Diffusion of Innovations. New York, NY, The Free Press, 1995, pp 252–280 26. Fallowfield L: Compliance issues in quality of life assessment: Experiences of two Cancer Research Campaign sponsored groups. Stat Med 17:541–546, 1998[CrossRef][Medline] 27. Hahn E, Webster K, Cella D, et al: Missing data in quality of life research in Eastern Cooperative Oncology Group (ECOG) clinical trials: Problems and solutions. Stat Med 17:547–559, 1998[CrossRef][Medline] 28. Kiebert G, Curran D, Aaronson N: Quality of life as an endpoint in EORTC clinical trials. Stat Med 17:561–569, 1998[CrossRef][Medline] 29. Olschewski M: Compliance with QOL assessment in multicentre German breast cancer trials. Stat Med 17:571–575, 1998[CrossRef][Medline] 30. Mosconi P, Torri V, Cifani S, et al: The multicentre assessment of quality of life: The interdisciplinary Group for Cancer Care Evaluation (GIVIO) experience in Italy. Stat Med 17:577–585, 1998[CrossRef][Medline] 31. Bernhard J, Peterson H, Coates A, et al: Quality of life assessment in International Breast Cancer Study Group (IBCSG) trials: Practical issues and factors associated with missing data. Stat Med 17:587–601, 1998[CrossRef][Medline] 32. Bernhard J, Gusset H, Hurny C: Practical issues in quality of life assessment in multicentre trials conducted by the Swiss Group for Clinical Cancer Research. Stat Med 17:633–639, 1998[CrossRef][Medline] 33. Osoba D, Zee B: Completion rates in health related quality of life assessment: Approach of the National Cancer Institute of Canada Clinical Trials Group. Stat Med 17:603–612, 1998[CrossRef][Medline] 34. Kaasa S, Hjermstad MJ, Jordhoy M, et al: Compliance in quality of life data: A Norwegian experience. Stat Med 17:623–632, 1998[CrossRef][Medline] 35. Moinpour C, Lovato L: Ensuring the quality of quality of life data: The Southwest Oncology Group experience. Stat Med 17:641–651, 1998[CrossRef][Medline] 36. Coates A, Gebski V: Quality of life studies of the Australian New Zealand Breast Cancer Trials Group: Approaches to missing data. Stat Med 17:533–540, 1998[CrossRef][Medline] Submitted November 9, 2001; accepted September 29, 2002.
This article has been cited by other articles:
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||
|
|||||||||||
|
|
Copyright © 2003 by the American Society of Clinical Oncology, Online ISSN: 1527-7755. Print ISSN: 0732-183X
|
TOP
ABSTRACT
INTRODUCTION
2 test for categorical variables and the Mann-Whitney U test and the independent samples two-sample Student’s t test (using Levene’s test for equality of variances) for continuous variables. Descriptive statistics were used to express patient attitude, and the 
