ePRO eDiary Benefits: Problems with Paper
More than 20 years of research and trial experience have demonstrated that paper diaries provide unreliable PRO data. eDiaries leverage configurable clinical trial data quality controls - not available on paper - to solve the problem of poor data quality, low data yield and delayed time to PRO data access.
The Four Types of Paper Patients
Dr. Stuart Donovan, of Knoll Pharmaceuticals, has provided PHT with an insightful representation of the four common types of patients you can expect with paper diaries:
- Perfect Patients answer all appropriate items logically and neatly, but there are no assurances that the diary was completed on time, per the study protocol, or by the authorized subject.
- Forgetful Patients simply neglect to complete the diary, and because paper data isn't provided in real time, it could take weeks or months to realize that no data is being provided.
- Selective Patients provide some information, but it presents complicated data analysis issues - especially if primary endpoint data is missing. This leads to the need for more patients than expected in order to draw reliable scientific conclusions.
- Enthusiastic Patients have terrific energy in filling out paper diaries - but the writing is illegible, the answers are illogical and there are often AEs in the margins. This data is unusable. Imagine how valuable this patient would be, however, using a LogPad.
Evidence of Faked Measurements in the Literature
Scientific research has consistently demonstrated that paper diaries make it difficult for subjects to comply with the protocol. The table below reviews several references which together indicate that approximately 25% of paper data is invented. As alarming as that figure is, it only accounts for fake data, and does not include data made unreliable by temporal uncertainty or error.
| Reference | Measures Counted (as % of Expected) | Measures Reported (as % Percent of Expected) | Faked (as % of Reported) | Number of Subjects |
|---|---|---|---|---|
| Jonasson G et al, Eur. Respir J 14, 1999 | 77% | 93% | 17% | 163 |
| Milgrom H et al, J of Allergy & Immunol 98, 1996 | 58% | 95% | 39% | 24 |
| Spector S et al, J Allergy Clin. Immunol 7, 1986 | 47% | 90% | 48% | 19 |
| Straka RJ et al, Pharmacotherapy 17, 1997 | 55% | 71% | 23% | 68 |
| Verschelden P et al, Eur Respir J 9, 1996 | 44% | 55% | 20% | 20 |
| Chemelik F et al, Annals of Allergy 73, 1994 | 85% | 100% | 15% | 20 |
| Simmons M et al, Chest 118, 2000 | Not reported | Not reported | 30% | 101 |
| Mazze R et al, American J of Medicine 77, 1984 | Not reported | Not reported | 40% | 14 |
| Weighted Avg. (% Faked x N)/N |
24.4% |
Learn More about Problems with Paper PROs
What other scientific evidence exists about poor (and fake) data using paper diaries? Are there any practical ways to make paper data more reliable? Contact PHT for answers to these questions or any others you may have.
