Introduction 1 2 4 4 9 10 10 4 11 Methods For the purposes of this systematic review of surveys, the term PDA is used synonymously to refer to any handheld device. Some examples include the following: Blackberry; Palm operating system devices, which include Palm Tungstens, Handspring Visor, and Sony Clie; and Pocket PC devices, which include the Compaq iPAQ and HP Jordana. Data Sources Surveys were identified as a subset selected from a broader systematic review examining all studies related to handheld devices in health care settings. Thus, initial search strategies and retrieved articles reflected this more extensive focus. This comprehensive literature search was conducted in consultation with an information specialist. The searched bibliographic databases covered both medical and engineering disciplines, including the following eight databases: Medline, Current Contents, Inspec, BA/RRM, Biotechnology, Biological Abstracts, EI Compendex, and EMBASE. The search was restricted to English-language literature published January 1993 (corresponding to the development of the first palm device) to February 2005. An updated search of Medline (PubMed) and EI Compendex (EI Village 2) was run near the project’s completion (January 30, 2006). Furthermore, the reference lists from included studies were examined in an effort to identify additional surveys not captured in the reference databases. In addition, surveys identified from Google searches and those known to the authors to have been conducted by private market research firms as well as physician groups were nominated for inclusion in our screening. Electronic Search Strategy all survey Appendix 1 Eligibility Criteria 
                    Surveys were included for this present review if they met the following initial criteria: related to an application in human health care and involved the use of a PDA device; contained original data; written in English (not including abstract or conference proceedings); published after 1993; and specifically reported handheld usage rates (prevalence of PDA use as a metric) in populations of health care professionals who were surveyed about the extent of their PDA use. Although conference proceedings were excluded, if deemed potentially relevant, a cross-check was conducted to see if there was an ensuing journal publication. A survey was not included if the handheld device being evaluated had undergone extensive custom modifications. A final set of unique references was identified and posted to the proprietary Web-based screening system SRS (Systematic Review Software). Selection Process Figure 1 Figure 1 Modified QUOROM Flow Chart for Identified PDA Usage Surveys  Data Abstraction The contents of each included survey were abstracted by one reviewer (CG), with an additional research assistant providing verification (TR). Analysis Appendix 2 12 13 Table 1 Included surveys       Year of Survey/Publication Author Prevalence of PDA Use Health Care Professional Group 1 1999/2000 18 15% (use in clinical work) Physicians 2 NS/2001 19 47% (use in clinical work) Specialists (Internists) 3 2001/2001 * 20 60% (use in practice) Physician Executives (organizational survey) 4 2001/2001 21 19.3% (use in clinical practice) Physicians & Specialists 5 2001/2001 22 26% (use in practice) Physicians 6 2001-2002/2002 23 38% (NS) Specialists (Pediatricians) 7 2000-2001/2002 * 24 67% (use in practice) Residents (Family Medicine) (organizational survey) 8 2001/2004 25 26.2% (office-based use) Physicians 9 2001/2004 26 33% (use at work or home) Pharmacists 10 2001-2002/2004 27 75% (use in practice) Medical Residents 11 2002/2002 2 27.9% (use in clinical practice) Physicians & Specialists 12 2002/2002 * 28 33% (use in physician offices) Physician Executives (organizational survey) 13 2002/2003 29 46% (use at medical institutions) Specialists, Medical Residents, & Fellows (Internists) 14 2002/2004 30 35% (use at work) Specialists (Pediatricians) 15 2002/2004 31 61% (use on an academic health science campus) Health Sciences Faculty & Medical Residents 16 2003/2003 32 32.9% (use in clinical practice) Physicians & Specialists 17 NS/2003 33 36% (use alone or in conjunction with log-card procedure in documenting) Medical Residents (Family Practice) 18 2003/2003 * 34 75% (carry & use PDAs) Physician Excutives (organizational survey) 19 2004/2005 14 57% (use regularly in a work week) Physicians, Specialists (Surgeons), & Medical Residents 20 2004/2005 35 91% own; 85% use on daily basis; 9% weekly; 215% monthly Specialists (Anestheologists) 21 2001/2005 * 36 64% of programs report “most or all” residents use for clinical purposes Medical Resident Programs (Emergency Medicine) (organizational survey) 22 NS/2005 37 67% (NS) Nurse Practitioners & Students 23 NS/2005 38 45.1% (own or use daily) Specialists (Radiologists)  NS/2004 † 17 Difficult to interpret the prevalence numbers among the resident respondents Medical Residents (Obstetrics & Gynecology)  2004/2005 † 15  Physicians, Specialists (various), & Medical Students 16 NS = not specified * † Results Figure 1 Table 1 Survey Characteristics The included surveys were published between 2000 and 2005, with survey data collected between 1999 and 2004. One survey had a four-year lag between data collection and publication, three surveys had a lag of three years, and three surveys had a lag of two years. We were unable to determine publication lag in four surveys as no data collection dates were provided. Surveys were from the United States (16), Canada (4), Australia (1), both the United States and Puerto Rico (1), and both the United States and Canada (1). Survey methodology reflected the following: self-administered questionnaires distributed solely by mail (11); telephone interviews (2); Web-based online surveys (4); and combined distribution by electronic or postal mail as determined by the recipient (4). Two studies did not report the methodology used. Response rates ranged from 5.7% to 92.6% across 13 of the included surveys; 10 surveys did not report such rates. PDA Use In presenting the results, we group the PDA users by type of health care provider and personal characteristics (eg, age). In terms of PDA use, physician specialists were surveyed exclusively in five surveys. Three surveys examined practicing physicians, three included physicians and specialists combined, two included medical residents exclusively, while two surveyed an amalgam of physicians, specialists, medical residents, and/or students. Three surveys targeted physician executives and organizational practice leaders. One survey was directed at directors of family practice residency programs, while a further survey targeting individual PDA use in emergency medicine resident programs was completed at the organizational level. In addition to physicians as users of technology, one survey targeted practicing hospital pharmacists and another targeted a national sample of nurse practitioner students and faculty. One survey included faculty and residents across several health science disciplines, including medicine, dentistry, nursing, public health, pharmacy, and applied health science. Figure 2 19 37 38 20 24 36 28 34 Figure 2 Range plots of PDA usage by health care providers (n = 17); middle points represent range medians  18 24 19 22 22 25 23 26 27 30 29 21 2 32 32 20 28 34 31 14 dependent 35 38 37 While PDA use has clearly increased since 1999, it appears as though only a handful of studies have examined the prevalence and usage patterns of such technology outside of physician groups. Furthermore, when comparing the included surveys in depth, distribution of use is not uniform across selected characteristics of surveyed health care professionals. Therefore, further subgroup analyses from the included surveys are provided below. Patterns of handheld use are also briefly examined. Patterns of PDA Usage Age 39 22 30 23 19 29 32 14 37 Students and Medical Residents 24 27 31 29 33 14 37 15 15 Gender 19 29 30 32 14 P 37 37 Family Physicians versus Specialists 32 2 21 32 14 Large and Hospital-Based Practices 22 22 30 14 25 Urban versus Rural Physicians 30 21 2 32 Professional Use Five surveys considered PDA use in both a professional and personal context; 17 studies exclusively captured professional use. One study reported general prevalence rates for PDA use among pediatricians; however, it did not specify if use was in clinical practice or outside of work. In order to discern professional use more closely, we explored administrative PDA uses versus direct use in clinical patient care. We found that of the surveys that concern PDA use within a health care setting, 17 of 23 studies (74%) reported use pertaining to administrative or organizational tasks, while 14 of 23 studies (61%) addressed PDA use in patient care. Billing and coding were the most frequently performed administrative PDA functions in 50% of the surveys reporting administrative uses. This was followed by 44% reporting calendar scheduling, 31% reporting Web and email access, 25% reporting address book use, and 25% stating use in charting patient details into an electronic health record. Other reported administrative tasks included the following: word processing, calculator, charge capture, procedure documentation, outpatient tracking, resident hours, telephone message tracking, general time management/personal organizer, patient referrals, procurement of supplies, patient census, order entry, dictation, and passwords and pins. In terms of patient care, access to drug information was reported in 93% of the surveys reporting clinical PDA use, while 50% reported prescribing, 43% stated accessing patient records, 43% described medical calculator use, and 36% indicated use in reference to laboratory values. Other reported clinical PDA uses included access to medical references, patient tracking and patient reminders, clinical decision pathways and managed care applications, telemedicine, and diagnostic imaging or radiology applications. Only one survey reported PDA use for patient education, and one referred to PDA use for research purposes. Discussion 40 Our grouped survey data suggest that there is little information on the PDA usage rates among nonphysician health care providers. However, collectively, these data suggest that use of handheld devices has become a subject that health care professionals need to know about. By systematically gathering this usage information, it is difficult to deny the prevalence of PDAs in health care. With this basic understanding of current handheld usage patterns, we need to consider the impact of this development of mobile handheld technology on both practice and research. 41 42 43 44 45 41 46 41 Figure 3 Figure 3 41  The increase in PDA adoption means a potential reduction in hardware and training costs when using handheld devices in the provision of care and in research. Because of the high probability that target health care professionals may already have a handheld device and will already know how to use one, the overall hardware purchase costs could be reduced, and the end user will not necessarily have to be trained from scratch. To date, use of PDAs in health care appears to have preceded extensive evaluative research. PDA adoption rates, already high, continue to be a moving mark with projections for rapid growth in the short term. By comparing handheld device diffusion to other health information innovations, and by placing PDA use within existing diffusion models, we are able to better predict the future of handheld growth in health care and therefore develop more timely and appropriate evaluative research to accompany such growth. 16 47 15 It is worthy to note that, with the exception of one survey focusing on nurse practitioner students, little mention was made in the surveys of PDA use by students across health care disciplines, including medicine. Several universities in Canada and the United States now mandate use of PDAs for medical undergraduate students and residency programs; therefore, it is assumed this could potentially affect prevalence rates. However, because none of the included surveys examined mandated use, we are unable to infer if this is responsible for recent increases. However, this raises an important issue to be considered in future studies related to students and rates of handheld adoption. To better understand the prevalence rates among the included surveys, it became important to categorize the drivers for PDA use as either professional or personal. We therefore attempted to discern what specific PDA tasks the respective health care professionals were performing. This was done by classifying, whenever possible, the use as administrative versus care. On the surface, it would appear that administrative and organizational tasks on a PDA exceed those related to patient care, perhaps signaling where the growth in adoption is most likely to occur. use In conclusion, physicians are increasingly accustomed to using a PDA, and, therefore, technology expertise will not likely be a barrier to deploying handheld applications. There is an urgent need to evaluate the effectiveness and efficiency of specific tasks using PDA technology (eg, implementation, searching, reference, data entry, reporting) to inform those persons developing and those using handheld applications. Furthermore, it is not clear why there is a paucity of evidence on the extent of adoption of PDAs by other health care providers: is it that they lag in the use of this technology or is it simply that they have yet to be studied? Limitations This review has a number of limitations. Issues around response bias and inability to draw causal inferences weaken survey methodology. It may be the case that those surveyed feel a stronger affinity to the survey sponsor, who has a greater interest in the questions asked, or are in complete disagreement with the topic at hand. This can skew results in difficult-to-measure ways. Quite possibly, the nonrespondents are the least committed (ie, nonusers of PDAs). As a result, the critical objective of drawing a true random sample of the populations that are the focus of the survey is compromised and the findings somewhat impure. The reported methodologies across these surveys appear to be heterogeneous, which limits their comparability. As noted, the quality of the included surveys could not be determined given the absence of validated quality assessment instruments, and, therefore, there was no adequate way to assess the influence of bias. A related issue is that some of the included surveys did not go through a rigorous peer-review process. These combined issues made judging the strength of the evidence not possible. One would assume surveys identified from scientific journals would be a source of less biased information. However, in defense of the nonacademic surveys, there is a consistency in results between those peer-reviewed versus those that were not. This may suggest that our main conclusions regarding adoption rates are fairly robust and not disconnected even with the inclusion of non–peer-reviewed evidence. Conclusions The objective of this study was to determine the adoption rates of PDAs in health care settings, and to project expected adoption in the future based on established technology diffusion models. Our findings from a systematic review indicate the current overall adoption rate for professional use of PDAs among health care providers, namely physicians, is 45% to 85%. Younger physicians, residents, and those working in large and hospital-based practices are more likely to use a PDA. Professional use in health care settings appears to be more focused on administrative tasks when compared to those related to patient care, although this requires further study. The adoption rate is now at its highest rate of increase according to a commonly accepted diffusion of innovations model. Additionally, the impact of PDA use on practice appears to be immediate in terms of costs and training. Familiarity will not likely be a barrier to deploying handheld applications in health care. However, there is a critical need to evaluate the effectiveness and efficiency of specific tasks using handheld technology within the health care system and across health care provider PDA user groups.