Background
The optimal use of medicines is a fundamental goal in health policy worldwide, especially given the growing pressure from chronic multi-morbidity in aging populations increasing the potential for inappropriate polypharmacy and negative ramifications for patient outcomes [
1,
2].
The safe and effective use of medicines relies on a large, constantly increasing knowledge base [
3]. Despite the current era of unprecedented access to information, there is evidence that unmet information needs remain an issue in clinical practice [
4]. Unmet needs are not simply due to a lack of information; rather, the new challenge is ‘information overload’ [
5]. Consequences of information overload include failure to process all available information, incorrect processing of information, accepting lower quality information, and abandoning the search for needed information [
6]. Information failures may contribute to less than optimal use of medicines or medicines-related harm [
7].
It is recognised that doctors regularly have information needs or questions relating to the care of their patients and that keeping up to date is a constant challenge [
8]. Research characterising information needs in primary care suggests that questions about medicines dominate [
9‐
11].
The main barriers identified in the literature preventing primary care physicians from meeting their general clinical information needs are lack of time and skills to complete a search and appraise information efficiently and effectively [
5,
12]. However, literature specifically addressing the challenges and barriers to meeting
medicines information needs in general practice is lacking. Research that focuses on the
medicines information-seeking behaviour of general practitioners
1 (GPs) is very limited and dated, though pharmacists and medicine compendiums have been identified as useful to assist in appropriate prescribing [
6]. Given health system emphasis on medicines optimisation in the face of chronic multi-morbidity and ongoing rapid evolution of the information environment, a clearer understanding of the present situation is important. The aim of this study was to identify the medicines information needs of GPs during routine clinical practice, the strategies they use to meet these needs, and the challenges they face. Improved understanding in these areas will help to inform enhancement of medicines information resources and support services for primary care
2 and therefore contribute positively to the optimal use of medicines as well as harm reduction.
Methods
A mixed methods approach with a convergent design [
13] was used to investigate GPs’ medicines information needs. This approach allowed the selection of methods best suited to answering the research questions. The two methods used were:
1.
Quantitative descriptive research using a structured reflection template to capture data about the medicines information needs arising and the information-seeking strategies used by GPs over two clinic days.
2.
Qualitative interpretive research using a follow-up face-to-face semi-structured interview to explore GP views on their medicines information needs including resources used, challenges experienced in practice and potential solutions.
Following approval by the University of Otago Human Ethics Committee (reference number D15/314) a purposive sampling strategy was used, recruiting GPs in active clinical practice from two areas of New Zealand: a metropolitan city and a smaller provincial city, and their surrounds. Flyers advertising the study were sent to the practice manager of all general practices in the provincial city (n = 28) and an equivalent number in the metropolitan city. Practice managers distributed the flyers to GPs within their practice. GPs contacted the research team if they were interested in participating. Informed consent was obtained from all participants. They were offered a voucher as a token of appreciation for their willingness to participate. This incentive was considered important to attain sufficient participant numbers recognising the time constraints of GPs and that they are a highly researched group [
14].
For the structured reflection, one of the authors (CC) met face-to face with each GP individually. During this meeting, GPs used the appointment list within their patient management system to verbally reflect on each consultation during the two preceding clinic days. They were prompted to describe the medicines information needs arising and resources used by questions from the researcher (see Additional file
1 for data collection template). To reduce the potential for memory recall bias, participants had also been provided with a documentation booklet (see Additional file
2) in advance of the two clinic days so they could make notes about the medicines information needs arising and resources used. During the structured reflection, key information was documented by the researcher using the data collection template. The meeting was also digitally recorded with participant consent and the audio recording used to verify and supplement data captured by the researcher. The structured reflection generated quantitative, descriptive data about the nature of medicines information needs arising and the resources used to meet them. This data was managed in Microsoft Excel and processed using descriptive statistics.
Structured reflection is an adapted version of post-consultation interview, a method used in previous research examining the information needs of clinicians [
6]. This approach enabled quantitative data about the medicines information questions arising in general practice to be gathered without the disruption to participants of being questioned after each individual consultation. It also reduced the potential for memory bias compared to survey [
15].
The semi-structured interviews were undertaken (also by CC) after the structured reflection, usually with an interval of 1–2 weeks. Some were undertaken immediately afterwards if that was more convenient for the participant. They were guided by an interview schedule (see Additional file
3) which focused on the nature of their medicines-related information needs, the resources they prefer to use, challenges experienced in practice, and potential solutions. These interviews were digitally recorded with participant consent, and transcribed verbatim. As an accuracy check, participants were given the opportunity to review their transcript and make amendments if necessary. The qualitative data were managed in NVivo (version 10, QSR International Pty Ltd) and an inductive thematic analysis undertaken [
16]. This involved iterative reading of transcripts and coding for potential themes and subthemes by CC. A second member of the research team (CM) read and independently coded six randomly selected transcripts. No new themes emerged during the latter interviews indicating that data saturation had been attained. The final thematic framework was discussed and agreed by all three authors. The analysis process considered both recurrent and outlying concepts emerging from the data. All members of the research team are pharmacists and have a professional interest in medicines information needs arising in healthcare and consider that unmet needs could hinder optimal outcomes from the use of medicines. Reflexivity was practised throughout the study with active consideration of the potential for researcher bias to influence study design and analysis [
17].
Integration occurred by comparison of the results obtained from separate qualitative and quantitative data collection and analysis. This approach enabled triangulation of the findings and took place during interpretation and development of the discussion [
13].
Discussion
This study provides contemporary insight into the needs and challenges that GPs have in finding information to support optimal medicines use. It confirms they have regular needs with their most common searches being about dose, interactions and side effects [
10,
18,
19]. It also suggests that outside this common core of medicines information needs, GPs are challenged by variation, complexity, ‘known unknowns’ and ‘unknown unknowns’. A deeper understanding of these most challenging elements of medicines use in general practice would enable information resources and support services to be tailored to better meet GP needs. A focus on gathering details of the ‘known unknowns’ encountered by GPs would allow assessment of whether they are truly unknown or whether increased searching time, different resources or different search strategies might help to reduce their occurrence. Knowledge gaps identified could then inform future research and strategies to support optimisation of medicines use [
20,
21].
The increasing challenge of complexity associated with multi-morbidity and polypharmacy contributed to time being a key issue for participants in this study [
22,
23]. When facing complex situations, the time and ability to locate, appraise, and interpret information about medicines to translate study findings into clinically meaningful information applicable to a specific patient is key. Medicines information services designed to provide timely patient-centred evidence-based clinical decision support are a potential option to help [
24,
25]. Indeed, in the present study, several participants in one area still used the hospital-based medicines information service they were aware of from previously working in the local hospital. As well as addressing the time constraints of general practitioners, medicines information services provide a human connection [
24]. This element may be important because a 2015 study of trainee GPs in Australia found that despite having trained in the ‘internet era’, human information sources were preferentially sought for more complex problems [
26]. Other research supports the existence of a preference for expert human resources in more complex scenarios [
27].
Although GPs in this study were positive about their relationship with local community pharmacists, they did not necessarily perceive them to be an accessible resource. There was limited awareness amongst participants of the support roles that pharmacists in different locations may play despite increasing literature indicating that general practice-based pharmacists can help to manage the growing burden of multi-morbidity and potential for inappropriate polypharmacy [
28‐
30]. This suggests room for improved inter-professional collaboration to address medicines information needs in primary care and improve the safe and effective use of medicines.
Time limitations affected resource use in various ways in this study. For example, there was no evidence of traditional primary literature-searching using databases such as Medline which are known to be time-consuming [
31]. Searching often transitioned from the core New Zealand-based medicines information references directly to Google, which is perceived as a quick route to wide-ranging information [
32].
The predominant use of New Zealand-based references identified in the current work aligns with a study on physicians use of resources for evidence-based medicine across three countries where a preference for tools and publications produced within their own country was observed [
33]. The authors of the previous study termed this a ‘cultural bias’, but for medicines information there are also practical trust-based drivers. Accurate local information is required regarding medicine availability, funding and regulatory status. Whether a medicine is approved by the local medicines regulatory authority may affect practitioner liability as well as the level of informed consent required for the patient.
This study identified that Google usage for medicines information needs occurs in general practice. The use of Google is not frequently mentioned in the wider literature but is consistent with one Australian study on internet use by GPs [
34]. It could be more common in primary care because subscriptions for specialised medical information resources routinely available in secondary care are less affordable for smaller organisations. The ready accessibility and ease of use are also likely to be key factors [
35,
36]. Google use warrants further examination given previous research suggesting that clinicians may be willing to accept lower quality or potentially biased information if it is available more quickly [
6,
32].
Trust was identified as a factor affecting use of resources for medicines information in the present study. In line with the work of Sim et al
., participants in this study often loosely described consideration of the trustworthiness of information retrieved from the internet [
34]. However, concerns have been raised in the literature about lack of systematic validation and the potential for hidden conflicts of interest with information sourced from the internet [
37,
38]. A study of emergency medicine residents reported less accurate answers to clinical questions when using internet information located via Google [
38]. Despite this, the residents were more confident in their answers when using Google. In a study of the online information searching of hospital physicians, Mikalef et al. also raised concerns about the use of non-authoritative information sources [
35]. A better understanding of how health professionals are using Google to meet their medicines information needs is needed.
Another finding of this study was the presence of the patient as a factor affecting GP medicines information-seeking strategies. Participants felt that patients generally accepted that doctors would need to search for information at times. However, there was an overall sensitivity to patient perception of in-consultation searching activities. Though this issue has not been widely explored, one early investigation of physician information-seeking behaviour suggested that clinician information needs may be suppressed due to embarrassment [
39]. A more recent study also found that some physicians preferred not to search for information during a consultation due to concerns that it would convey uncertainty or lack of knowledge [
36]. However, a 2011 investigation of perceptions of in-consultation information-seeking found that most patients do not lose confidence in their physicians as a result. It was also suggested that physicians, especially those early in their career, tend to overestimate the potential for loss of patient confidence [
40].
Our findings suggest the use of Google by doctors on the doctor-patient relationship warrants further attention. Google was perceived by participants as a useful information tool, but they were generally wary that it would be considered unprofessional by patients. Several mentioned that patients had commented to them with disapproval about other doctor’s use of Google. There is minimal research in this area though it has been suggested that internet search engine use by doctors may result in decreased patient confidence and a perception of reduced quality of care [
40]. The need for a similarly accessible and usable portal but limited to quality medical information has been raised in the literature [
35,
41].
The finding of significant drug interaction alert fatigue is consistent with international reports that alerts with low clinical relevance cause considerable clinician frustration and dissatisfaction [
42]. Given the context of unprecedented, growing levels of polypharmacy [
43] and the findings of this study that drug interactions are among the most common information needs expressed, this is of major concern and needs coordinated action by policy makers and software vendors.
The most frequently used resources in this study were accessed electronically via a desktop computer. This reflects the ease of access for GPs whose consultation workflow revolves around their computer, where they also document patient notes and generate prescriptions. In contrast with older studies [
44], recent literature is increasingly reporting electronic resources as dominant tools in the information-seeking of medical students and qualified doctors including GPs [
36,
45]. Despite the preference for electronic resources observed, smart phones were not used at all for medicines information retrieval by participants in this study. This is likely because there is no accessibility advantage to smart phones for GPs because their workflow revolves around a desktop computer. This suggests that when developing tools for GPs, the focus should remain on desktop applications rather than smart phone technology.
The mixed method approach used in this study provides a depth of understanding of the medicines information needs and challenges in general practice that could not be attained by either quantitative or qualitative findings alone. While some aspects of the nature of medicines information needs will vary with delivery of GP care in different countries, the factors influencing medicines information resource use identified in this study are likely to be relevant to GPs’ clinical practice in Western English-speaking countries with similar health care systems. Aspects contributing to the trustworthiness of the qualitative findings included diversity within the sample, participants reviewing their interview transcript to confirm accuracy, and attention to views that contradicted those of the majority during analysis.
A potential limitation for this study is the risk of sample bias. It is possible that GPs agreeing to take part had an interest in medicines information that was in some way different to those not taking part. There is also a risk of memory recall bias with self-report of information needs [
6,
8]. Attempts were made to mitigate this effect by provision of a documentation booklet and only asking participants to reflect on consultations within the previous two days. Another limitation is that the quantitative data is primarily descriptive.
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.