Symptom recording in a randomised clinical trial: paper diaries vs. electronic or telephone data capture☆,☆☆
Introduction
Patient-Reported Outcomes (PROs) represent an umbrella term that captures in a variety of ways how patients perceive symptoms or a disease, and how they respond to treatment. One of the most frequently used PROs is the measurement of frequency and severity of symptoms. Subjective assessments of symptoms offer a perspective that extends beyond biomedical markers and should be regarded as complimentary to the “objective” signs and symptoms of a disease [1]. Other commonly used PROs in conjunction with the evaluation of treatment efficacy are medication outcome and health-related quality of life assessment. PROs can be administered in different ways, with paper copies or electronic devices being two commonly used routes of data capture [2]. While these instruments have the advantage of being patient-reported, there are problems inherent in their use. Because paper-based instruments are unsupervised, errors may occur. Patients may misunderstand questions, miss questions, select more than one response option where one is intended, or place a mark between the options. Moreover, patients may be asked to register an assessment every morning and every evening for a period of time. A problem associated with this kind of registration is that patients do not always fill in the diary at the appropriate time. In fact, it is likely that patients may complete their diaries just before attending a study visit at the clinic. This has implications on memory retrieval of symptom occurrence and intensity, i.e., fluctuations during the week might be missed, speed of onset might not be recorded correctly, and the last impressions of either adequate or inadequate symptom resolution might be overestimated unless recorded on a daily basis. Additionally, there may be a further reason for bias related to the fact that human beings have a tendency not to be willing to disclose that the instructions have not been followed [3].
We believe that these findings have far-reaching implications. When patients in paper-based diary studies report high adherence, it must seriously be considered the possibility that actual adherence may have been much lower. In the absence of an objective, documentation of good actual adherence with paper-based diaries jeopardizes validity and statistical power, and might introduce bias because patients might report symptoms as perceived during the last days rather than on a daily basis.
There is an abundance of evidence indicating that memory is not a unitary phenomenon [4], [5] when doing subjective assessments. It is as much a matter of reconstruction as retrieval [6].
What is called episodic memory does very well illustrate this problem [7]. It has been shown that retrieval is unreliable and rife with inaccuracies and biases. This undermines the entire purpose of collecting diary data and the validity of paper diaries may therefore be questioned. If adherence can be improved, quality will be improved and money saved. The results of clinical trials would be strengthened, and the discontinuation rate reduced [8].
The objective of this study was to compare the accuracy, responsiveness, and autocorrelation of consecutive entries in paper diaries (P-Diaries) with electronic diaries (E-Diaries) and telephone diaries (T-Diaries) for the registration of symptoms through 4 weeks of treatment of gastro-oesophageal reflux disease (GORD). Additional objectives were to identify potential problems associated with the patients' use of the diaries and to compare assessments made by the investigator and by the patient.
Section snippets
Patients and study design
No statistical determination of sample size has been done since the study is of exploratory nature. However, approximately 50 evaluable patients were judged to be sufficient to give an estimate for detection of a difference in symptom assessments in diaries. In this multi-centre, open, randomised, and parallel study, a total of 177 patients (86 males and 91 females) were included in 13 centres in Denmark between March and August 2002. Nine centres were general practice and four were specialist
Results
All analyses were performed using the statistical software package SAS version 8.1 (SAS Institute, Cary, NC).
Discussion
To verify the proportion of entries completed on time and the proportion of entries completed half a day late, we compared P-Diaries with E-Diaries and T-Diaries. As noted, for P-Diaries entries could be made outside the designated time. Our study demonstrates that the number of entries that were done on time or half a day late was relatively high in each of the diary groups, with highest proportion for E-Diaries. A probable explanation for the low proportion of entries in T-Diary might be that
Acknowledgments
We thank all participating investigators who provided and cared for study patients and A. Malm who conducted all the statistical analyses.
References (20)
Compliance with treatment and the art of medicine
Am. J. Cardiol.
(2001)- et al.
Cross-cultural adaptation of health-related quality of life measures: literature review and proposed guidelines
J. Clin. Epidemiol.
(1993) - et al.
Should study subjects see their previous responses?
J. Chronic. Dis.
(1985) Long-term aims of treatment of reflux diseases, and the role of non-drug measures
Digestion
(1992)- et al.
Ecological momentary assessment: real world, real-time measurement of patient experience
- et al.
Patient non-compliance with paper diaries
BMJ
(2002) - et al.
The Betula Prospective Cohort Study: memory, health, and aging
Aging Neuropsychol. Cogn.
(1997) - et al.
Defining cognitive changes in depression and anxiety: a psychobiological analysis
Psychiatr. Psychobiol.
(1988) Real-time self-report of momentary states in the natural environment: computerized ecological momentary assessment
- et al.
Recall biases and cognitive errors in retrospective self-reports: a call for momentary assessments
Cited by (61)
The use of patient reported outcome measures (PROMs) in palliative radiotherapy: A topical review
2020, Radiotherapy and OncologyMenstrual questionnaires for clinical and research use
2017, Best Practice and Research: Clinical Obstetrics and GynaecologyCitation Excerpt :Both diaries and questionnaires have been used to collect data on patient-based outcomes among women with HMB. Prospective electronic data collection is increasingly being used to study medical conditions in their “natural environment,” when multiple data entries and the specific times of the entries must be obtained [56–58]. This method is also known as ecological momentary assessment, experience sampling, or real-time data collection [59,60].
Challenges during long-term follow-up of ICU patients with and without chronic disease
2016, Australian Critical CareCitation Excerpt :Use of electronic diaries can improve the quality of data captured but in our case was not available for the follow-up of patients. Advances in this technology have shown it to be successful in improving compliance with study protocols and in the accuracy of information provided by patients.32,33 Researchers of other follow-up studies planning to use electronic diaries need to ensure patients are comfortable with information technology such as smart phone applications or a web-based method of recording information.
Influence of psychological factors on acute exacerbation of tension-type headache: Investigation by ecological momentary assessment
2015, Journal of Psychosomatic ResearchCitation Excerpt :This was an advantage over EMA that uses paper-and-pencil diaries [16] as in previous studies, particularly when investigating temporal relationships. Furthermore, “faked compliance,” which is a disguise of compliance by recording data at times other than those designated, has been discussed with respect to paper-and-pencil diaries [16] and faked compliance not only makes the time axis inaccurate but also may affect the data itself [10]. Such data inaccuracy can also be avoided in cEMA.
- ☆
Contributors: K. Lauritsen was the principal investigator of the study, revised the clinical protocol, assessed patients, interpreted data, and wrote the final manuscript as well as the early versions. A. degl' Innocenti was the project leader of the study, created the hypothesis and the design for the study, wrote and edited the clinical protocol, analysed and interpreted the data, and wrote the final manuscript as well as the early versions. L. Hendel, J. Præst, and M.F. Lytje assessed patients and revised the protocol as well as all the versions of the manuscript. K. Clemmensen-Rotne was the study coordinator, revised the clinical protocol, obtained data, provided advice on the fieldwork, and revised all versions of the manuscript as well as wrote the final manuscript. I. Wiklund created the design for the study, wrote the grant proposal, edited the clinical protocol, and analyzed and interpreted the data as well as edited the manuscript.
- ☆☆
Conflicts of Interest Statement: K. Lauritsen, L. Hendel, J. Præst, and M.F. Lytje received reimbursement from AstraZeneca for costs related to the conduct of the study. A. degl' Innocenti, K. Clemmensen-Rotne, and I. Wiklund were employed at AstraZeneca.