Clinical trials are the most expensive and time consuming steps in drug development. In a paper-based clinical trial process, a Case Report Form (CRF), which is data collection form, typically has three or four copies per page, and the monitoring is typically done by hand at the investigative site comparing the CRF with source documents. The CRF data are later entered twice at a central location into a database. Information technology has been used to support clinical trials to have the most immediate and noticeable impact on drug development. Clinical trial technology that has been applied to improve clinical trial data collections is generally referred to as electronic data capture (EDC). The benefits of EDC include direct data entry at the investigator site leading to greater accuracy, fewer queries, decreased paper record storage, and timelier population of the study database1). Although increasing in pace, the conversion to EDC has been a slow progression2). A survey conducted in the year 2001 reported that only 5 percent of all new clinical trials made use of remote data entry for clinical trials, whereas the remaining 95 percent continued to use paper3).

The internet allows transfer of clinical trial information and coordination of multiple clinical trial processes. Therefore, the Internet have increasingly been utilized in clinical trials4)5). However, wired internet needs to connect the computer to the location where the Internet connection is available. Recently, mobile computing technology is proposed to improve information access, enhance workflow, and promote evidence-based practice to make informed and effective decisions at any place and any time. Mobile devices have been used in clinical trials for more than 16 years, utilizing different technologies. These devices have been used in different medical fields such as diabetes6)7), nutrition8), different surveys9). The mobile technology capable of data capture or data entry becomes employed in clinical trials. The variety of applications in clinical trials ranges from electronic patient diaries10-12) to electronic case report forms (eCRFs)13)14).

Personal Digital Assistants (PDAs) are one the most popular mobile devices embedding mobile internet communication network. The PDAs offer portable and unobtrusive access to clinical data and relevant information at the point of care. They have been extensively tested and have shown mixed results as a data collection tool by physician users15)16) and patient users17) based on the application. The main advantage of PDA is to link a PDA to a central computer anytime and anyplace, whereas the wired systems are strained to the location where the connection exists. It enables to centralize study information and coordinate multiple trial processes in real time at a lower cost18). In Korea, mobile internet subscribers exceeded 30 million, about 70% of the population by 200219). The mobile internet is generally used data communication network for everyday life such as mobile shopping, mobile banking, and mobile advertising in Korea. In spite of spreading mobile technology in Korea, data collection in clinical trial studies has still done on paper first and sometimes with electronic data entry taking place later. Furthermore, even few PDA-based clinical data collection systems typically collect and store data in offline mode, and then synchronize with the server by transferring to the operational database. Also, data review, data quality checks, and monitoring were performed manually and locally at the site. Real time data collection and transmit through mobile internet connection is still challenging area.

In response to these challenges, we have developed a wireless PDA application that patients can use both to evaluate their experience with a drug and to directly transmit their data to a database using a wireless internet connection. We used PDA instead of paper patient diaries in a randomized, placebo-controlled, double-blind irritable bowel syndrome (IBS) trial. IBS, a functional gastrointestinal disorder, rely exclusively on patients' self reports of their symptoms20).

The purpose of this study is to determine the improvement of speed, data accuracy and protocol compliance of PDA clinical data collection system comparing with traditional paper based data collection, which is popular way of current study, in the irritable bowel syndrome clinical trials study.

A randomized double-blind study was carried out to verify whether Remeron compared to placebo relieves irritable bowel syndrome (IBS). The study was approved by the Institutional Review Board of the hospital, and was designed as a double-blind clinical trial for efficacy of Remeron for IBS patients. Initially, sixty-four IBS patients from an 800-bed hospital signed informed consent forms. Subjects were required to visit one week before treatment for screening, to determine whether they met the inclusion criteria. Subjects took the actual medication or the placebo medication for four weeks and submitted a questionnaire each week. Patients visited the site each week during the one week of screening period and at weeks 2, 3, 4, 5, and 6 of the treatment period (Fig. 1). The study was conducted from Nov. 2001 to Apr. 2003.

The sixty four patients, who agreed to participate in the IBS clinical trial for the Remeron, were self-selected to either the experimental (PDA data entry) group or the control (standard paper data entry) group, according to their preference. Subjects of both control group and intervention group are asked to complete the questionnaires consisting of 6 check lists everyday as suggested in Table 1.

Research coordinators gained the informed consent and enter patients' basic information into the system at the beginning. Both forms are completed once daily at home for five weeks.

The eighteen subjects in the control group are requested to fill in the questionnaires daily on the paper CRFs for one week and bring the completed papers to next planned visit.

On the other hand, the eighteen subjects in the intervention group are requested to answer questionnaires on the PDA till two o'clock of the next day of taking medication everyday and transmit data to the server through wireless internet.

This result in 630 data sets consisting of six questionnaires each, and 3,570 data points total. The intervention group subjects complete the six items on the PDA and transfer the complete data it by two o'clock each day using wireless internet. The control group subjects were asked to fill out the paper daily, and submit seven paper CRFs each week at their planned visits. Then, a coordinator entered their data into a web application later.

Measurement instruments for the effectiveness of the PDAs in collecting clinical trial data were developed in consultation with the co-authors of clinical side,: a) response rate; b) completeness of data; c) the number of input errors; and d) protocol compliance. Response rate measured the extent to which they used the e-CRFs during the treatment period. Response completeness was measured whether the subjects answered all data items completely. Matthew et al. also proposed completeness of data to evaluate data quality for PDA data collection21). A chi-square test was conducted to compare response rates and response completeness between the two groups. Additionally, the Mantel-Haenszel chisquare test was applied to identify the influence of data collection methods on response completeness controlling demographic variables. Finally, the number of input errors and protocol compliance were measured.

Protocol compliance means that both groups should enter their medication intakes during five weeks at five measurement periods on a paper or a PDA form. Subjects usually report that they fill out the survey form following the protocol. However, subjects usually forget to write down the form at the exact time. This study compared the reported compliance and actual compliance.

Sixty four patients initially enrolled to participate, but twenty eight patients withdrew for the following reasons suggested by Table 2.

This study addresses whether wireless PDA technology can reduce clinical data input error and improve response rate, data quality and protocol compliance.

Average age choosing PDA group is much lower than non-PDA group. Higher preference for PDA of younger generation is likely due to their equal familiarity with mobile phone. As mobile phones continue to become integrated into their daily lives, mobile EDC preference among patients may continue to grow.

In spite of participants' preference, the response rate was lower than paper group. We assume that current PDA technology is not stable and reliable to use for clinical trials yet based on our supplementary survey. The survey regarding the PDA usage shows that few users become disgruntled and complained some difficulties of PDA data entry. Out of 21 respondents, 66% of them showed the difficulty of PDA usages, and 33% had unstable internet access. To apply PDA for clinical data collection, these technical issues should be resolved soon.

Our finding also supports previous literature reporting that PDA can enhance completeness of data collection21). Lower number of the missed items was found in the PDA collection compared with paper method. This increased data collection has a direct beneficial impact on the quality of data collected.

Regarding to input errors, PDA is supposed to have error free data collection compared with paper. However, PDA collection still has input errors. The errors were functional errors, like pushing the wrong button or the same button twice. Therefore, predesigned functions alerting users before developing e-CRF application will be able to reduce possible errors. Electronic data entry will not result in error-free data collection without built in check functions. Minimizing the data entry errors that result from extra data entry steps will streamline the data clean-up stage of clinical trials.

In spite of few errors, the quality of data in PDA is improved in comparison to paper group because the data are more complete, have fewer errors, and there are fewer protocol violations like other studies14)22-24). Thus, data editing, retyping, proofreading, and the process of clarification can be eliminated, and irrelevant questions can be omitted.

Also, protocol compliance was comparably improved. Significant retrospective entries were made in the control group. Poor diary completion may result from having unreasonable expectations of patients and giving incomplete instructions. Electronic, time coded diaries could ensure better quality of records.

Selection bias can occur because the intervention group is self-selected according to their preference. In addition, recruitment only included patients treated by an investigator in large hospital. This sampling method limits the degree of generalization of our results and may not reflect outcomes in other setting. Additional attempt to control potential impact of selection bias should be made in this type of research. Also, this study should be replicated using a larger number of subjects from various clinical trial studies in future.

This study shows advantages and some weakness of wireless PDA technology applying for clinical trials. Our results showed input errors are significantly reduced but PDA based data collection cannot improve response rate. PDAs with wireless internet capabilities have the potential to improve clinical trial data collection and management, and thus, to expedite the development of important new drugs. In spite of various overseas researches in the area of electronic diary, PDA has not been used in clinical trial yet in Korea. However, wide adoption of PDA technology is expected with better designed PDA hardware and software applications, faster internet access, improved battery life, seamless integration of PDA technology with hospital information systems, and satisfactory security measures. Herein, we described some barriers and benefits based on our experience to collect clinical trial data. We desire it will give you some idea on how to successfully implement PDA based data capture system in your organization. The results of this study confirm the feasibility and adaptability of the data capture through PDA administered questionnaires. Analysis of data completeness, response rate, and protocol compliance suggest that the IBS questionnaires can be maintained under PDA application.