As depicted in Figure 1, 54 doctors (43 residents and 11 interns) owned a PDA. Forty-five doctors (83.3%) used a Palm operating system while 9 doctors (16.7%) used Windows CE. The majority (33 doctors, 61.1%) used their PDAs for both personal and work-related purposes, as compared to 12 (22.2%) with mainly personal purposes and 9 (16.7%) with mainly work-related purposes.

With respect to using the PDA to acquire medical knowledge, the median time spent per week was 0.5 hour (IQR 0 – 2.0 hours). The median frequency of use per week was 4.0 times (IQR 0 – 12.3 times). The median score on a Likert scale of 1 to 5 (where a score of 1 represents "definitely useful" and 5 represents "definitely not useful") for the perceived usefulness of the PDA for acquiring medical knowledge was 2.0 (IQR 1.0 – 2.0). A detailed breakdown of how these doctors rated its usefulness is provided in Figure 1.

These PDA owners perceived that the PDA was most useful for providing drug information, followed by medical references, scheduling and the calendar, medical calculators, and finally documentation (Table 2). However, in reality, half or less of these same owners had the appropriate medical software applications for drug information, medical references and medical calculators downloaded into their PDAs (Table 2). In total, the following programmes were downloaded into the 54 PDAs: drug information: the Sanford Guide to Antimicrobial Therapy (7 programmes), Epocrates (4), Physicians' Desk Reference (4), Johns Hopkins POC-IT Antibiotic Guide (2), British National Formulary (1), MIMS (1), Mosby's Drug Consult (1), Royal Children's Hospital Drug Doses (1); medical references and guidelines: 5-Minute Consult Series (26), Harrison's Manual of Medicine (12), Washington Manual of Medical Therapeutics (12), the ICU Book (4), Handbook of Evidence-based Critical Care (3), American College of Cardiology Guidelines (2), International Classification of Diseases – Ninth Revision (1), eMedicine (1), Nelson Textbook of Pediatrics (1), Schwartz's Principles of Surgery (1), Williams Obstetrics (1), Dahnert's Radiology Review Manual (1), ARTBeat (1), .911 (1); medical calculators: Archimedes (8) and other medical calculators (12).

As for potential drawbacks of the PDA, the perceptions of all 134 respondents on these are listed in Table 3. The greatest concerns highlighted were the fear of loss and breakage, and the preference for working with desktop computers and paper. Only a small minority felt that the PDA made one look unprofessional, was too technical, or worked too slowly.

Of the remaining 80 doctors who did not own a PDA, only 14 (17.5%) thought they would buy one in the next 1 year (Figure 1).

As depicted in Figure 2, 76 doctors (58 residents and 18 interns) had previously used UpToDate. The median time spent per week was 1.0 hour (IQR 0.5 – 2.0 hours). The median frequency of use per week was 3.0 times (IQR 2.0 – 5.0 times). The median score on a Likert scale of 1 to 5 for the perceived usefulness of UpToDate was 1.0 (IQR 1.0 – 1.0). The median time taken to find an answer per search was 5.0 minutes (IQR 3.0 – 10.0 minutes).

Most of these 76 doctors felt that hospitals should have institutional subscriptions to UpToDate, and most would recommend UpToDate to a colleague (Table 4). However, most would not subscribe personally to UpToDate even if the hospital ceased its institutional subscription. They were also less enthusiastic in pronouncing that UpToDate had led to a change in their diagnoses or management of patients, or that UpToDate helped to avoid referrals to other specialties (Table 4). Among the 58 residents who used UpToDate, about half had used it in their clinics. Indeed, among the users of UpToDate, residents used it more frequently than interns (median 3.0 times per week [IQR 2.0–5.3 times] versus 1.5 times [IQR 1.0–3.5 times], p = 0.004, Mann-Whitney U test). Table 4 also shows the features of UpToDate which the doctors felt made it popular.

Only 93 doctors (69.4%) were aware that our hospital had an institutional subscription to UpToDate. Of these, 4 doctors (3.0%) had a personal subscription to UpToDate. Trainees in the fields of internal medicine, paediatrics, obstetrics and gynaecology, and family medicine – specialties which are featured in UpToDate – were more likely to be aware that the hospital had an institutional subscription to UpToDate (90.0% versus 65.8%, p = 0.03, chi-square test), more likely to have used UpToDate previously (85.0% versus 51.8%, p = 0.006, chi-square test), more likely to find UpToDate useful (70.0% versus 45.6%, p = 0.04, chi-square test), and spent more time using UpToDate (median 1.0 hour [IQR 0.4–2.0 hours] versus median 0.25 hours [IQR 0–1.0 hours], p = 0.009, Mann-Whitney U test) compared to other doctors.

There were 23 doctors (17.2%) who used both UpToDate and PDAs downloaded with medical software. Among them, although the same amount of time was spent per week on both tools, PDAs were used more frequently than UpToDate. Nevertheless, UpToDate was perceived to be more useful than the medical software applications downloaded on the 23 PDAs for acquiring medical knowledge (Table 5).

Internationally, the popularity of the PDA is rapidly increasing among doctors 910. Systematic reviews of multiple studies have found that between 45% and 85% of doctors use PDAs 1415, and that physicians in training were more likely to use one than more experienced doctors 15. In our study, 40.3% of our house staff owned a PDA, a rate which is obviously lower than that found in other studies. Even among the owners of PDAs, the median time spent using the PDA per week was only 0.5 hours. This is despite most owners of PDAs reporting use for both work and personal purposes. In addition, only a small minority of those who did not own a PDA would want to buy one in the next 1 year. Aside from the fear of breakage and loss – concerns which are prevalent among PDA users 1617 – many doctors preferred working with desktops and paper, suggesting that the advantage of portability of PDAs was not sufficient to persuade most doctors to make the switch to using PDAs.

Further analysis of our results yields the necessary information to explain the relative under-use of PDAs by these doctors. Most PDA owners in our survey perceived that the PDA provided useful drug information, medical references, scheduling functions and medical calculators. These are the same features which have been highlighted in other medical studies on PDAs 14151618192021222324. However, although these functions were rated highly, only half of our PDA owners had medical references downloaded onto their PDAs (interestingly, none of the owners had UpToDate downloaded on their PDAs 1213), and only one-third or less had medical calculators and drug information software applications downloaded. In previous studies, residents and medical students who were provided with PDAs which were already loaded with the relevant medical software applications usually rated their experience with the PDAs highly 252627. In all, this suggests that in order to make the PDA more effective as an information tool for house staff, residency programmes should strongly consider the provision of the necessary subscription for PDA medical software applications, along with technical support and the installation of these applications 24.

With regard to UpToDate, more than 90% of all users in our study agreed that the following features made it popular: the synthesis of relevant information, being updated regularly, its comprehensive references, its wide collection of subspecialties, and its easy-to-use recommendations. Given that UpToDate features topics in internal medicine, paediatrics, obstetrics and gynaecology and family medicine, it was understandably more popular among trainees within these disciplines. Indeed, prior studies that had demonstrated the popularity of UpToDate had mostly evaluated doctors or medical students within these disciplines 111162829.

Our study however revealed two concerns on UpToDate. First, although UpToDate was generally perceived to be useful, only slightly more than half of the house staff had used UpToDate. Although our institutional subscription began 5 months before the study, 30.6% of residents and interns did not know about it. This implies that more work needs to be done to inform our doctors of the resources available in the hospital. Second, although 86.8% of UpToDate users found the resource useful, only slightly more than half of them felt that it had led to a change of management in their practice, and only slightly less than half of them felt that it had led to a change of diagnosis and a decrease in the amount of referrals to other specialties. This may reflect a lack of confidence or ability to integrate new information into actual day-to-day medicine. Senior doctors should be mindful of this and build a working environment which encourages learning and the adoption of new clinical practices 33031.

The use patterns of the 23 doctors who have used both medical software applications on PDAs and UpToDate nicely sum up some of the features of these 2 tools. Although PDAs were used more frequently, more time was spent on UpToDate. This highlights the portable nature and accessibility of PDAs 910 – which may be used for quick information retrieval by busy house staff – as opposed to UpToDate, which provides textbook-like information, usually on desktop computers, at a more leisurely pace 12. It must be emphasised that our survey did not compare the usefulness of UpToDate versus PDAs per se. Rather, we found that the medical software applications downloaded on our doctors' PDAs was perceived to be less useful than UpToDate for retrieving medical information. This may again reflect a lack of technical support for doctors using PDAs.

The limitations of our study should be acknowledged. First, since the study was based in a tertiary care hospital in Singapore, the results may not be generalisable to community-based hospitals and residency training programmes, and to the developing world. Second, though the response rate to our questionnaire was reasonably good at 79.8%, we do not have sufficient information on the non-respondents to determine the presence of a response bias. Third, as this is a questionnaire study, the validity of our results depends on how the respondents' answers truly reflected their actual practice. Also, the questions in the survey were generated by us and have not previously been validated. Fourth, we limited the questionnaire to the assessment of the use of PDAs to acquire medical knowledge and did not evaluate the integration of PDAs with administrative work processes and clinical information systems in the hospital 9101415. Fifth, while we directly compared the time spent and perceived usefulness of PDAs and UpToDate, given their very different functions, our questionnaire could not provide a direct comparison of their various pros and cons. Sixth, various advances in both tools have been made since the time of our survey. For example, more topics are now covered by UpToDate, and refinements have been made to its search engine. Meanwhile, the distinction between PDAs and phones is being increasingly blurred with the rapid emergence of smart phones with PDA functionality 910. Importantly, although none of our house staff's PDAs contained software for UpToDate, at the time of our survey, UpToDate could actually be downloaded onto PDAs with a Windows CE system but not a Palm operating system (nevertheless, Now, not only can UpToDate may be downloaded onto most PDAs, faster wireless access has also facilitated the use of UpToDate on PDAs and smart phones with web browsing capabilities.