Skip to main content
NCBI home page
Wiley Open Access Collection logoLink to Wiley Open Access Collection
. 2025 Feb 24;91(8):2151–2161. doi: 10.1002/bcp.70009

Understanding factors that influence the drug choice of prescribers: A Q‐methodology study

Mariëlle G Hartjes 1,2,3,✉, Annelot E F Elsevier 1,2, Jan Willem Grijpma 4,5, Milan C Richir 1,2, Michiel A van Agtmael 1,2, Jelle Tichelaar 1,2,3
PMCID: PMC12304866  PMID: 39993935

Abstract

Introduction

Therapeutic decision‐making often involves weighing several treatment options on the basis of, among other things, the disease profile, patient characteristics but also prescriber‐specific factors. This study investigated the factors influencing therapeutic decision‐making among prescribers and explores how these factors differ between healthcare professionals.

Methods

This Q‐methodology study involved 37 participants from different medical backgrounds, including physicians, dentists, midwives, physician assistants and advanced nurse practitioners from various intra‐ and extramural settings in the Netherlands. Participants ranked 55 factors on their importance for medication prescribing, and factor analysis was used to identify distinct prescriber profiles. These profiles were subsequently enriched using qualitative data derived from post‐sorting interviews.

Results

Four prescriber profiles were identified that explained 59% of the study variance: pragmatic contextual, guideline oriented, experience driven and vulnerability focused. Several factors were ranked differently in each profile. The pragmatic contextual profile emphasized patient context and risk prevention, the guideline‐oriented profile adhered to clinical guidelines, the experience‐driven profile relied on clinical experience and patient‐centred communication, and the vulnerability‐focused profile prioritized personalized care for specific patient groups.

Conclusion

This study sheds light on the diverse factors influencing prescriber decision‐making in medical practice. By identifying four prescriber profiles, it reveals the complexity and diversity in how prescribers balance guidelines, clinical experience and patient‐specific considerations in their practice. These findings offer direction for developing educational and policy initiatives to enhance the quality of prescribing, ultimately improving patient outcomes.

Keywords: education, medical education, prescribing, quality use of medicines

1. INTRODUCTION

The prescribing process is susceptible to errors, 1 , 2 , 3 , 4 , 5 errors which may result in patient harm, diminished quality of life and increased healthcare costs. 6 , 7 While several strategies have been introduced, such as harmonizing European clinical pharmacology and therapeutics (CP&T) education and placing a greater focus on skills‐based training, they have not succeeded in reducing prescribing errors substantially. 8 , 9 A better understanding of the reasoning processes underlying prescribing may help improve prescribing practice.

Choosing the best treatment for a patient involves more than following guidelines: it requires consideration of various disease‐, patient‐ and medication‐related factors. Even then, prescribers with similar patients may opt for different therapies. 10 The current understanding of how these prescribing decisions are made is limited and largely drawn from studies of cognitive psychology and diagnostic reasoning. These studies often distinguish between intuitive, non‐analytical thinking (type 1) and analytical thinking (type 2). 11 Analytical thinking involves weighing various treatment options based on different factors, with patient‐related, practice‐related, drug‐related, information‐related and teaching‐related factors being considered crucial. However, how prescribers ultimately make decisions, and which factors they prioritize, remains unclear.

In the Netherlands, prescriptive authority, formerly limited to physicians, dentists and midwifes, has been broadened to physician assistants (PAs) and advanced nurse practitioners (ANPs). This makes understanding decision‐making processes more complex because each profession has its own values and perspectives. Although prescriptions from different healthcare providers may be comparable, there may be differences in the therapeutic decision‐making processes involved. 12 , 13 Factors such as personal opinion, experience, clinical setting and professional background may influence prescriber's reasoning strategies and priorities. 14 , 15 , 16 , 17 For instance, treatment emphasis may vary between settings, with some prioritizing cure while others may prioritize care/quality of life. Moreover, junior doctors may copy their supervisors' treatment approaches, often based on clinical experience, including known drug‐related factors. 18

Given the complexity of therapeutic decision‐making and the involvement of diverse prescribers, it is important to understand their approach and the factors they consider essential to prescribing. The aim of this study was to identify the factors that prescribers consider important in therapeutic decision‐making and if and how these factors differ between prescribers. This knowledge can be used to develop comprehensive, reflective and engaging education programmes for prescribers, potentially improving their prescribing skills in daily practice.

2. METHODS

2.1. Q‐methodology

This study used Q‐methodology, a mixed‐method research technique for investigating subjectivity. This methodology was developed in 1935 by William Stephenson and has since been used in healthcare and medical education studies. 19 , 20 , 21 It is a person‐centred analysis technique that examines the relationships between variables across subgroups, ultimately identifying distinct subgroups of individuals. 22 The methodology consists of five steps: (1) Q‐set development, (2) participant selection, (3) data collection, (4) data analysis and (5) factor interpretation.

2.1.1. Step 1: Q‐set development

The Q‐set is a set of statements related to the research topic. To create a balanced and representative set of statements, the Q‐set was revised on the basis of four rounds of feedback. The initial Q‐set consisted of 81 statements derived from relevant literature and prescribers' experience. Then four pharmacotherapy experts from different professions and with different levels of experience reviewed the Q‐set for clarity and completeness during a think‐aloud procedure. Unclear statements were either deleted or rephrased to improve clarity and five new statements were incorporated. Next, the research team discussed the Q‐set to assess its relevance, wording, overlap and completeness. Then a larger group of pharmacotherapy experts reviewed the Q‐set for consensus on agreement and wording. Lastly, the Q‐set was tested in a pilot session with two prescribers with average knowledge and interest in pharmacotherapy and six pharmacotherapy experts to evaluate its relevance and clearness. The final Q‐set consisted of 55 statements (Table 1), which falls within the recommended range of 40‐60 statements. 22 This process is shown in Figure 1.

TABLE 1.

Q‐set statements and factor arrays.

No. Statement P1 P2 P3 P4 A
1. Patient's age 1 1 ‐2 3 1
2. Disease severity 2 5 2 0 2
3. Use of national treatment guidelines 3 3 1 1 2
4. My own clinical experience 2 0 4 2 3
5. Available dosage forms of a medication −1 −1 0 1 −1
6. Pharmacokinetic characteristics of the medication 0 1 −1 1 0
7. Potential of a medication to reduce the patient's symptoms 1 4 5 5 5
8. Minimizing side effects and toxicity 2 0 2 1 1
9. Likelihood of resistance (with prolonged treatment) 1 1 −1 −3 −1
10. Need for medication monitoring −1 −3 −2 −4 −3
11. Negative experience with a medication 0 0 0 −1 0
12. Reports in the media about a medication (positive or negative) −5 −2 −4 −1 −4
13. Negative consequences of the patient's condition on quality of life 1 −1 2 4 2
14. Patient's adherence to therapy 0 −1 0 1 0
15. Patient's medication allergies 5 5 3 0 5
16. Medication dosing frequency 0 0 0 −1 −1
17. Environmental effects of a medication −3 −4 −5 −5 −5
18. Received continuing education on this topic −1 2 −1 −3 −1
19. Local prescribing agreements 0 −1 0 0 0
20. Assessment of the patient's wishes based on their socio‐cultural background −3 −2 −1 0 −2
21. Potential that a medication can cure disease 0 0 3 −4 0
22. Potential interactions of a medication with other drugs, intoxications and food 3 3 1 3 3
23. Patient's risk factors 2 4 0 1 1
24. Patient's wishes 2 −2 4 5 4
25. Colleagues' experience and knowledge 0 1 1 −1 0
26. Patient's renal function 5 −2 2 0 2
27. Patient's comorbidities 4 2 0 2 3
28. Education received on this topic during training −2 1 −1 −1 −1
29. Potential of a medication to improve quality of life 3 0 5 3 4
30. The administrative burden of prescribing −5 −5 −3 −3 −5
31. Patient's other medications 4 2 −1 1 1
32. Evidence‐based effectiveness of a medication 1 4 2 3 3
33. Patient's weight 1 3 −2 −1 0
34. Use of international guidelines −1 1 0 2 0
35. Availability of a medication −3 −1 −2 2 −2
36. Earlier course of the disease 0 2 0 4 1
37. Medication‐induced prolongation of life expectancy 0 −5 0 −2 −2
38. Costs of a medication −3 −3 −2 0 −2
39. Patient's ethnic background −2 −3 −4 −2 −3
40. Information provided by the pharmaceutical industry −4 −1 −3 −2 −3
41. Willingness to try different, equivalent medications to gain experience with as many medications as possible −4 −4 −3 −4 −4
42. Expected side effects of a medication 3 2 1 2 2
43. Positive experiences with a medication 0 0 3 0 1
44. Potential of a medication to prevent recurrence −1 −1 4 −1 0
45. Number of therapeutic indications for a medication −2 0 −2 −3 −2
46. Patient's gender −2 −3 −5 −5 −4
47. Availability of new medications −4 −2 −3 −2 −3
48. Patient's life expectancy 2 −4 1 0 0
49. Feedback from other healthcare professionals about choice of medication −2 0 1 −2 −1
50. Availability of both long‐acting and short‐acting forms of a medication −1 −2 −1 0 −1
51. Potential of a medication to slow disease progression 0 0 3 0 1
52. Examples of clinical supervisors/mentors −1 2 0 0 0
53. Patient's vulnerability 4 0 1 4 4
54. Justified deviation from guidelines if more appropriate for the patient 1 3 2 2 2
55. Potential interactions of a medication with dietary supplements or natural remedies −2 1 −4 −2 −2
Open in a new tab

Abbreviations: A, second order analysis a; P1, profile 1; P2, profile 2; P3, profile 3; P4, profile 4.

FIGURE 1.

FIGURE 1

Open in a new tab

Q‐set development.

2.1.2. Step 2: Participant selection

Strategic sampling was used to select prescribers with broad prescriptive authority in the Netherlands. A broad prescriptive authority was chosen because these professionals encounter a wide range of issues, which requires complex reasoning skills. In contrast, a limited prescriptive authority, such as that held by diabetes nurses, involves a restricted selection of medications, which may lead to different considerations. Participants from all over the Netherlands were selected from different specialties and professions, with different levels of experience, ranging from novices to specialists. Inclusion criteria included regular prescribing duties, appropriate for the professional group and broad prescriptive authority in the Netherlands. Large sample sizes are not required in Q‐methodology because inverted factor analysis is used and does not aim for population generalization. It is recommended to include the same number of participants as statements or slightly fewer. 22 As such, the aim was to include 40 prescribers in this study.

2.1.3. Step 3: Data collection

Participants underwent a 1‐h interview, conducted either face‐to‐face or online through Microsoft Teams, according to their preference. This made it possible for participants from all over the Netherlands to participate. The interview consisted of three sections. Initially, participants filled out a sociodemographic questionnaire using Castor EDC. 23 They then ranked statements from the Q‐set based on the perceived importance of these statements to therapeutic decision‐making (Q‐sorting). They were given both verbal and written instructions before sorting, emphasizing the relative nature of the ranking and encouraging responses based on real‐life practice rather than ideal scenarios. Participants began by categorizing the Q‐set into three preliminary groups (most important, least important and neutral) using either printed cards or online via EQ‐WebSort. 24 They then placed the statements on the Q‐grid, which had a fixed, quasi‐normal distribution, ranging from −5 (least important) to +5 (most important), as depicted in Figure 2. Once sorting was complete, they photographed the final arrangement. Subsequently, in a post‐sorting semi‐structured interview, the interviewer delved into the rationale behind participants' choices, their observed behaviour during sorting and their insight into their decision‐making process. The interview questions can be found in Appendix A. The interview was recorded and provided qualitative data that were used to assess the factor solution and enrich factor interpretations.

FIGURE 2.

FIGURE 2

Open in a new tab

Q‐sorting grid, where participants can sort statements by importance.

2.1.4. Step 4: Data analysis

The Q‐sorts were analysed using a by‐person factor analysis (principal components factor analysis with varimax rotation) with Ken‐Q Analysis Desktop Edition (KADE version 1.3.0.), which is specialized software for by‐person factor analyses in Q‐methodology studies. 24

In the first step, factors were included based on three absolute criteria: (1) an eigenvalue of 1.0 or higher according to the Kaiser–Guttman criterion. 22 In Q‐methodology, an eigenvalue represents the amount of variance in the data that is accounted for by a particular factor, indicating its strength. Higher eigenvalues often indicate dominant or widely shared perspectives among study participants, while lower eigenvalues reflect less prevalent but still meaningful viewpoints; (2) all the factors combined should explain at least 35–40% of the variance in total, with at least two Q‐set sorts loading significantly on a factor. The study variance refers to the total variability in the dataset that is being explained by the factors and (3) the use of relative criteria, for example if the qualitative interview data supported the factors and if the factors made sense (coherent, differentiated, recognizable). M.H., A.E., J.W.G. and J.T. evaluated the outcomes and determined the accepted solution through consensus. For factor loading, a significance cut‐off point of 0.45 was used. This cut‐off value was chosen to obtain distinctive profiles that included as many people as possible. A value higher than 0.44 was not allowed for a second factor (confounded factor loading). In that case, the participant was not included in a profile. Because of the high correlation between the profiles, the first‐order profiles were submitted to a second‐order analysis to highlight the similarities, as suggested by Watts and Stenner. 22 s‐order analysis has been used for many years in factor analysis and Q‐methodology. 25 , 26 , 27 , 28 The factor scores from the first‐order factors were added to KADE and a new analysis, principal component without rotation, was made based on those profiles. This led to factor analysis of the original profiles to determine the consensus between them.

2.1.5. Step 5: Factor interpretation

Factor interpretation was performed to understand the perspective encapsulated by the factor and shared by the participants. As proposed by Watts and Stenner, 22 the process began with the calculation of factor arrays, which represent the weighted averages of the Q‐set sorts within a factor. These arrays revealed how a typical prescriber associated with a factor would arrange the statements. Interpretation entailed analysing the statements given the highest and lowest rankings initially, followed by those showing significant deviation and those ranked in the middle. Both quantitative and qualitative data contributed to this process, providing a comprehensive depiction of each factor.

2.2. Ethical aspects

The Medical Ethics Review Committee of the Amsterdam University Medical Center determined that the study did not fall within the scope of the Medical Research Involving Human Subjects Act (WMO) (reference: 2023.0645). Informed consent was obtained in advance from all participants.

3. RESULTS

3.1. Participant characteristics

Thirty‐eight participants, physicians, dentists, midwives, physician assistants and advanced nurse practitioners, participated. One participant was excluded for not completing the study. To ensure a wide range of perspectives, participants were selected to maximize variance across several demographic factors. For instance, the study included a balanced representation of gender, workplace settings (hospitals, primary care, private practices and community health centres) and departments. Participants came from eight of the 12 provinces in the Netherlands. The demographic data are presented in Table 2.

TABLE 2.

Demographic information.

Sample Profile 1 Profile 2 Profile 3 Profile 4
Number of participants 37 14 5 8 4
Age in years (mean (range)) 40.6 (29‐61) 40.0 (29‐54) 38.0 (29‐47) 40.8 (29‐56) 47,8 (32‐61)
Gender 19 (51.4%)
Female 18 (48.6%) 6 (42.9%) 4 (80%) 4 (50%) 3 (75%)
Male 8 (57.1%) 1 (20%) 4 (50%) 1 (25%)
Experience in years (median (range)) 7 (0‐33) 6 (0‐22) 8 (4‐17) 12 (0‐20) 7.5 (5‐10)
Professions
Physician 16 8 ‐ 4 1
PA 8 4 ‐ 4 ‐
ANP 8 2 1 ‐ 3
Dentist 4 ‐ 3 ‐ ‐
Midwife 1 ‐ 1 ‐ ‐
Specialties
Intramural 16 5 1 7 2
Surgical 7 2 ‐ 5 ‐
Non‐surgical 9 3 1 2 2
Extramural 21 9 4 1 2
Care by appointment 14 6 4 ‐ 1
Care at personal residence 7 3 ‐ 1 1
Open in a new tab

Abbreviations: ANP, advanced nurse practitioner; PA, physician assistant.

3.2. Factor analysis and profile description

On the basis of the absolute and relative criteria, four distinct prescribing profiles were identified. A factor represents a shared viewpoint from a particular group of prescribers (from now on called “profiles”). Each profile was characterized by different decision‐making processes and factors that influenced the prescribers' approach to patient care, highlighting the diversity in prescribing practice. These profiles provide insight into how various prescribers prioritize patient‐specific factors, guidelines, clinical experience and individual patient needs, among other things, in their prescribing practice. The profile descriptions below are the result of the factor interpretation process. The statement number and its position on the distribution grid are given in brackets. These statements are supported by extra information provided by participants in the post‐sorting interview. An overview of the core values of the four profiles can be found in Table 3 and a graphical overview is presented in Figure 3.

TABLE 3.

Overview of the four different profiles.

Profile 1: Pragmatic contextual Profile 2: Guideline‐oriented Profile 3:Experience‐driven Profile 4: Vulnerability‐focused
Approach Pragmatic, evidence‐based, focused on utility and usability Guideline‐oriented, with flexibility when necessary Experience‐driven, focused on quality of life Vulnerability‐focused, personalized care
Focus Patient context, risk prevention Safe therapy, symptom reduction, patient preferences Experience, feedback, quality‐of‐life considerations Individual needs, off‐label use if necessary, less emphasis on complete cure
Key considerations Renal function, allergies, comorbidity, comedication Risk assessment: disease severity, allergies, risk factors Reduction of symptoms and disease progression, patient wishes Patient wishes, symptom reduction, impact of disease on quality of life
Open in a new tab

FIGURE 3.

FIGURE 3

Open in a new tab

Graphical overview of the four different profiles and their overlap.

3.2.1. Profile 1: Pragmatic contextual prescribers

Factor 1 had an eigenvalue of 15.939 and explained 43% of the study variance. Fourteen participants were significantly associated with this profile. They were predominantly employed extramurally in general practice or residential settings (such as nursing homes or rehabilitation care) and had varying levels of experience. Four participants with relatively little experience worked in hospitals.

Prescribers in this profile use an evidence‐based approach based on their use of national guidelines (S3, +3), but the individual patient in front of them is even more important. The patient is viewed in context, and especially his/her vulnerability (S53, +4). Vulnerability encompassed physical and mental aspects, as well as various risk groups. With a view to reducing the risk of adverse events, prescribers focus on patient‐specific factors such as renal function (S26, +5), allergies (S15, +5), comorbidity (S27, +4) and comedication (S31, +4). The most critical consideration is an absolute prescribing contraindication, as they cannot be ignored. Since the medication has to be acceptable to the patient, expected side effects (S42, +3) are taken into account. Statements about the medication made in the media (S12, −5) are important for patient discussions but do not influence therapy choices. Administrative burden is perceived as bothersome, but is considered a general problem in healthcare rather than an impediment to patient care. The availability of new treatments (S47, −4) is taken into consideration less often in primary care settings, where many participants worked, than in a hospital setting.

In conclusion, prescribers in this profile adopt a pragmatic approach. This approach is often evidence‐based and responsive to practical considerations, such as usability for the patient in their own context. The focus is on the patient's context, including relevant patient‐specific factors, with emphasis on preventing serious risks.

3.2.2. Profile 2: Guideline‐oriented prescribers

Factor 2 had an eigenvalue of 2.2999 and explained 6% of the study variance. Five participants were significantly associated with this profile. They worked primarily extramurally, mainly “by appointment”.

Prescribers in this profile indicate that prescribing is often supportive of their primary treatment, such as dental treatment or non‐pharmacological advice. They typically prescribe a limited number of medication groups, which make them experts in a small range of medications. They assert that guidelines form the basis of their prescribing practice, with guidelines founded on effectiveness as reported in the literature (S32, +4). The reasons for strictly following guidelines vary. Some view it as a professional responsibility, while others see it as adhering to hospital policy. A reason for deviating from guidelines involves assessment of disease severity (S2, +5), allergies (S15, +5) and risk factors (S23, +4). Patient weight (S33, +3) is also crucial, especially since paediatric dosages often need to be calculated based on weight. Patient preference (S24, −2) plays a minor role. Participants note that patients often request antibiotics without an indication. In such cases, they adhere to guidelines rather than acquiesce to these requests. However, a realistic patient request can be a reason to deviate from guidelines. Prescribers prioritized symptom reduction (S7, +4); prolonging life expectancy (S37, −5) was not a major concern because it was not relevant in their field. Similarly, a patient's life expectancy (S38, −4) is less relevant, as these prescribers often worked with relatively healthy patients.

In conclusion, prescribers in this profile are guideline oriented. Guidelines are central to their approach, although deviations can be made when necessary. The options given in a guideline are evaluated based on various factors, both patient‐and medication oriented, with emphasis on selecting a safe therapy.

3.2.3. Profile 3: Experience‐driven prescribers

Factor 3 had an eigenvalue of 1.9536 and explained 5% of the study variance. Eight were significantly associated with this profile, five of whom had different levels of experience and worked in surgery departments and two of whom were experienced prescribers and worked in internal medicine departments.

Prescribers in this profile indicate that the primary goal is to improve quality of life (S29, +5) by reducing symptoms (S7, +5), slowing disease progression (S51, +3) and preventing symptoms in the future (recurrence of complaints) (S44, +3). This is cited as the primary objective of a prescriber. They aim to do what is best based on their own knowledge and experience (S4, +4), and particularly rely on previous positive experiences (S43, +3), which make them more likely to repeat the same actions. Their experience is also formed by feedback from colleagues and supervisors, guidelines and conferences. Their knowledge and experience are either broader or more limited depending on where they worked. If their experience is insufficient, they might look for additional information or refer patients. Experience is combined with patient preferences (S24, +4). By treating the symptoms, it is assumed that they are meeting the patient's wishes, as they assume that symptom management is often a part of these wishes. It is considered important not to ignore patient wishes. Like experience, patient wishes are based on multiple aspects, such as past experiences of the patient or their background. These prescribers consider cure (S21, +3) to be relatively more important than did prescribers in other prescribing profiles. Although cure is seen as ultimate goal, this is not always possible, therefore prescribers in this profile prioritize quality of life. There are certain factors they regularly lack knowledge about, making it difficult to take them into account, such as patient's gender (S46, −5), ethnic background (S39, −4) or environmental effects of medication (S17, −5).

In conclusion, prescribers in this profile are experience driven. Quality of life is central to these prescribers, and they make use of various factors, both patient‐ and medication‐oriented, when making prescribing decisions, with emphasis on their own clinical experience.

3.2.4. Profile 4: Vulnerability‐focused prescribers

Factor 4 had an eigenvalue of 1.794 and explained 5% of the study variance. Four participants were significantly associated with this profile. They predominantly worked with specific groups of patients, such as the frail elderly, children or patients receiving palliative care.

The patient as individual is central for prescribers in this profile, which also includes the vulnerability of the patient (S53, +4). The focus is on what is desirable and tolerable for the patient, as discussed with the patient. Important factors here are patient wishes (S24, +5) and the negative impact of the disease on quality of life (S13, +4). To improve the quality of life (S29, +3), prescribers aim to reduce symptoms (S7, +5) and prevent or slow disease progression (S36, +4). The achievement of complete cure (S21, −4) is not always applicable to their patients, who often have chronic disease or received palliative care. In other acute and bridgeable settings, this can be different, but it is generally not the primary goal. The patient's age (S1, +3) also plays a role, as it is related to life expectancy, and influenced the goals of treatment or the route of drug administration (eg, in children). The need to monitor the patient (S10, −4) was not a major factor in medication choice because this task would often be delegated to others. Prescribers often felt they lacked sufficient knowledge of some factors to apply them effectively, such as the environmental impact of medicines (S17, −5) and the role of patient (S46, −5), to influence their prescribing choices.

In conclusion, prescribers in this profile are vulnerability focused. Quality of life is central to this profile, partly because these prescribers work with vulnerable patients. Given their work with these specific patient groups, they often cannot rely on standard guidelines or there may not be relevant guidelines, which often necessitates the off‐label use of medications. The emphasis is on personalized prescribing and consideration of the patient's individual context.

3.3. Second‐order analysis

First‐order profiles 1, 2, 3, and 4 were highly correlated, indicating a degree of consensus among them. Therefore, a second‐order analysis was performed, leading to second‐order profile A (see Table 1), which highlights the underlying consensus in the four first‐order profiles. This resulted in an overarching common viewpoint on which individuals from all profiles agreed.

The most important factors were the potential of a medication to reduce the patient's symptoms (S7, +5), patient's medication allergies (S15, +5), potential of a medication to improve quality of life (S29, +4), patient's wishes (S24, +4) and patient's vulnerability (S53, +4). Conversely, the least important factors were the environmental effects of a medication (S17, −5), the administrative burden of prescribing (S30, −5), the willingness to try different, equivalent medications to gain experience with as many medications as possible (S41, −4), the patient's gender (S46, −4) and reports in the media about medication (S12, −4).

Aspects of all four first‐order profiles can be found in this common viewpoint: reducing patient's symptoms (profiles 3 and 4), the patient's medication allergies (profiles 1 and 2), the potential to improve quality of life (profile 3) and the patient's vulnerability (profiles 1 and 4). This shared viewpoint is characterized by two main goals, namely, reducing symptoms and improving quality of life. These goals can be achieved by taking factors such as allergies, the patient's wishes and their vulnerability into account. Other factors can be helpful, but vary more per profile. There was a negative consensus regarding the role of the environmental effects of medication and the patient's gender, mostly because prescribers lacked knowledge of these factors. Participants considered that the administrative burden of prescribing should not stand in the way of reaching treatment goals and is therefore considered least important.

4. DISCUSSION

This study investigated what factors prescribers consider important in therapeutic decision‐making and if and how these factors differ between prescribers. Four distinct prescriber profiles were identified, reflecting the multifaceted nature of prescribing in clinical practice. Although there were significant differences (Table 3), there were also common therapeutic goals, such as the aim to reduce symptoms and improve quality of life. This highlights the shared core values of healthcare professionals, but the approaches to achieve these goals can vary by profile.

Previous research has emphasized the complexity of clinical decision‐making, noting that it involves a combination of evidence‐based guidelines, clinical judgement, patient preferences and other factors. 29 Our findings are consistent with this, showing that prescribers integrate and weigh various sources of knowledge and factors when prescribing. Although this is the first study to rank these factors, their importance has long been recognized. For instance, the emphasis on guidelines in profile 2 is consistent with the findings of Grimshaw et al, 30 who found that adherence to clinical guidelines can improve patient outcomes. Similarly, the pragmatic approach seen in profile 1 is consistent with research by Sinsky et al 31 that highlights the importance of considering individual patient contexts in primary care. The experience‐driven approach of profile 3 is consistent with the findings of Bensing 32 and Gabbay and Le May, 33 which emphasize the role of experiential knowledge and patient‐centred communication in effective medical practice. The focus on patient vulnerability seen in profile 4 underscores the importance of personalized care for specific patient groups, as highlighted in the work of Ekman et al 34 and Robinson et al. 35 In the past, the literature has often focused on single aspects, whereas this study ranked multiple factors involved in decision‐making, providing a good representation of the complexity in daily practice.

4.1. Strengths and limitations

A strength of this study is that this is the first use of Q‐methodology to rank factors important to prescribing. This approach provides new insights into the decision‐making process and the role of subjectivity in this. This is important because little is known about how prescribers integrate various factors to reach a final decision. Weighing different factors to make a decision is part of analytical, type 2 thinking. Another strength of this study is the inclusion of various groups of healthcare professionals with prescriptive authority in the Netherlands, unlike previous research that often focused primarily on single professions. This comprehensive approach provides a more complete picture of prescribing practice across the healthcare sector.

A limitation is whether all possible perspectives were captured and whether the same perspectives would have been captured if different choices had been made during participant selection. However, it is important to be aware that this is inherent to the method chosen. Nevertheless, careful consideration was given to selecting participants from throughout the Netherlands. During the selection process, extra attention was given to identifying which perspectives might be missing from the initial selection of participants, thereby ensuring a diverse range of viewpoints and the transferability of findings. The number of midwives in the study was limited, but is not expected to significantly impact the results because midwives have a more restricted prescribing role, similar to that of dentists. We found that midwives and dentists all met profile 2, therefore it can be expected that adding extra participants would not reveal more viewpoints and thus change the profiles. Additionally, it is important to note that the four profiles identified in this study are specific to Dutch clinical practice, which, for example, is characterized by prescribing rights that are not limited to doctors. The country‐specific context may have influenced the profiles that emerged, therefore it is possible that different profiles could be identified in other countries or contexts.

Owing to the complexity of the medication prescribing process and Q‐sorting, participants found it challenging to weigh the importance of various factors. Opinions on the “most important” factors were often clear and strong, as participants consciously reflected on these factors. For factors that participants were less aware or conscious of, opinions tended to be weaker. Unconscious factors are more likely part of their type 1 reaction, which might be analytically reconsidered by conscious factors in type 2 thinking. 36 Because type 1 thinking occurs unconsciously, it can be difficult to rank the items because prescribers are not aware of this process. Factors that were placed in the “least important” category were frequently those that did not apply to the participants or about which they had insufficient knowledge. This indicates either a believed or known lack of importance or a knowledge gap regarding certain factors, such as gender or environment, and how to incorporate these into the therapeutic reasoning process.

4.2. Implications for medical education

This study identified four different prescriber profiles among Dutch healthcare professionals. Based on these findings, some prudent recommendations can be made, although more research is necessary. Since these four profiles are common among prescribers and there are many factors involved in decision making, prescribers need to have the appropriate tools to make optimal decisions based on their own values and viewpoint. It is important that during their training, students are aware of these profiles and how they affect the decision‐making process. Therefore, several suggestions can be made to improve CP&T teaching. Firstly, certain topics, such as gender and environment, may need more emphasis in education 37 , 38 because participants mentioned that they would use these factors more frequently if they had the necessary knowledge. Students need to gain a solid understanding of these factors and how to apply them. For example, case studies that explore different routes of administration and the impact of medicines on the environment can provide valuable opportunities for discussion. Secondly, teaching programmes should include these four prescriber profiles and the factors that are relevant to them. This would ensure that prescribers have insight into factors that are important in the medication prescribing process in daily practice. Case studies for students may need to be diversified to include vulnerable patients, quality of life and patient wishes.

4.3. Future research

The findings of this study can serve as a foundation for future research. Given the low generalizability of the Q‐methodology because of the limited number of participants, a follow‐up survey study might be helpful to learn more about the prevalence of the different profiles. 39 It is hereby important to look at an even more varied group of prescribers to also represent the views that may currently be underexposed. A follow‐up survey will help ensure that the findings apply to a broader range of prescribers. Additionally, these profiles are based on how prescribers typically behave in their usual settings. However, it would be interesting to explore whether these behaviours evolve over time and if they are context‐specific. If they are indeed context‐specific, it could be valuable to investigate whether prescribers can be trained to consciously select the most appropriate profile for different situations. Consideration must also be given to how these different prescriber profiles could most effectively be addressed in education to ultimately improve daily practice. It would be interesting to see whether students from the different professions are characterized by specific profiles. This could help to either emphasize or weaken specific therapeutic reasoning aspects in their training to prescribe.

5. CONCLUSIONS

This study provides a nuanced understanding of the diverse factors influencing prescriber decision‐making in medical practice. The identification of four distinct prescriber profiles, namely, pragmatic contextual, guideline oriented, experience driven and vulnerability focused, highlights the complexity and variability in how prescribers integrate guidelines, clinical experience, patient‐specific factors, etc in their practice. These insights can guide educational and policy interventions aimed at improving the quality and consistency of prescribing practice, ultimately improving patient outcomes.

AUTHOR CONTRIBUTIONS

M.H., A.E. and J.T.: Conceptualization, writing of the original draft. J.W.G.: Conceptualization, analysis, revising. M.R. and M.vA.: Conceptualizing, revising, and editing. All the authors approved the final version.

CONFLICT OF INTEREST STATEMENT

The authors declared no competing interests for this work.

ACKNOWLEDGMENTS

We would like to thank the participants in our pilot session and our participants in this study. We would like to thank Rachel Kreinz for the creative development of the figure.

Appendix A. Post‐sorting questions

Questions used during the post‐sorting interview:

  1. Why did you place these statements at the extremes (both positive and negative)?

  2. Are there any statements in the middle that you would like to comment on?

  3. Did you miss any statements while sorting?

  4. Discussion of notable points during sorting.

  5. Are there statements that, in an ideal situation, you would place elsewhere? If so, which ones and why?

  6. You have indicated how you currently make decisions. What are the reasons that might lead you to suddenly decide a different way?

  7. How do you combine all the different inputs at the time of prescribing?

  8. To what extent does the setting (clinic or outpatient clinic) influence your choice? Is the location/population also relevant in this context?

  9. Did you make the sorting with a specific patient in mind or perhaps a clinic visit or an entire month?

  10. Do you easily prescribe medication or are you hesitant about it?

  11. How would you describe your role as a prescriber?

Hartjes MG, Elsevier AEF, Grijpma JW, Richir MC, van Agtmael MA, Tichelaar J. Understanding factors that influence the drug choice of prescribers: A Q‐methodology study. Br J Clin Pharmacol. 2025;91(8):2151‐2161. doi: 10.1002/bcp.70009

The authors confirm that the Principal Investigator for this paper is Jelle Tichelaar. No patients were involved in this study.

Funding information The authors did not receive funding for this project.

DATA AVAILABILITY STATEMENT

The data that support the findings of this study are available from the corresponding author upon reasonable request.

REFERENCES

  • 1. Leendertse AJ, Egberts AC, Stoker LJ, van den Bemt PM. Frequency of and risk factors for preventable medication‐related hospital admissions in the Netherlands. Arch Intern Med. 2008;168(17):1890‐1896. doi: 10.1001/archinternmed.2008.3 [DOI] [PubMed] [Google Scholar]
  • 2. Lewis PJ, Dornan T, Taylor D, Tully MP, Wass V, Ashcroft DM. Prevalence, incidence and nature of prescribing errors in hospital inpatients: a systematic review. Drug Saf. 2009;32(5):379‐389. doi: 10.2165/00002018-200932050-00002 [DOI] [PubMed] [Google Scholar]
  • 3. Slight SP, Howard R, Ghaleb M, Barber N, Franklin BD, Avery AJ. The causes of prescribing errors in English general practices: a qualitative study. Br J Gen Pract. 2013;63(615):e713‐e720. doi: 10.3399/bjgp13X673739 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 4. Avery AJ, Ghaleb M, Barber N, et al. The prevalence and nature of prescribing and monitoring errors in English general practice: a retrospective case note review. Br J Gen Pract. 2013;63(613):e543‐e553. doi: 10.3399/bjgp13X670679 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 5. Dean B, Schachter M, Vincent C, Barber N. Causes of prescribing errors in hospital inpatients: a prospective study. Lancet. 2002;359(9315):1373‐1378. doi: 10.1016/S0140-6736(02)08350-2 [DOI] [PubMed] [Google Scholar]
  • 6. Dornan T AD Heathfield H Lewis P, et al. An in‐depth investigation into causes of prescribing errors by foundation trainees in relation to their medical education: equip study General Medical Council; 2009. [Google Scholar]
  • 7. De Vries TP, Henning RH, Hogerzeil HV, Fresle DA. Guide to good prescribing: a practical manual 1994. [Google Scholar]
  • 8. Brinkman DJ, Tichelaar J, Okorie M, et al. Pharmacology and therapeutics education in the European Union needs harmonization and modernization: a cross‐sectional survey among 185 medical schools in 27 countries. Clin Pharmacol Ther. 2017;102(5):815‐822. doi: 10.1002/cpt.682 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 9. Donker EM, Brinkman DJ, van Rosse F, et al. Do we become better prescribers after graduation: a 1‐year international follow‐up study among junior doctors. Br J Clin Pharmacol. 2022;88(12):5218‐5226. doi: 10.1111/bcp.15443 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 10. Tichelaar J, Richir MC, Garner S, Hogerzeil H, de Vries T. WHO guide to good prescribing is 25 years old: quo vadis? Eur J Clin Pharmacol. 2020;76(4):507‐513. doi: 10.1007/s00228-019-02823-w [DOI] [PubMed] [Google Scholar]
  • 11. Kahneman D. Thinking, fast and slow. Farrar, Straus and Giroux; 2011:499. [Google Scholar]
  • 12. Jiao S, Murimi IB, Stafford RS, Mojtabai R, Alexander GC. Quality of prescribing by physicians, nurse practitioners, and physician assistants in the United States. Pharmacotherapy. 2018;38(4):417‐427. doi: 10.1002/phar.2095 [DOI] [PubMed] [Google Scholar]
  • 13. Stewart D, MacLure K, George J. Educating nonmedical prescribers. Br J Clin Pharmacol. 2012;74(4):662‐667. doi: 10.1111/j.1365-2125.2012.04204.x [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 14. Neufeld VR, Norman GR, Feightner JW, Barrows HS. Clinical problem‐solving by medical students: a cross‐sectional and longitudinal analysis. Med Educ. 1981;15(5):315‐322. doi: 10.1111/j.1365-2923.1981.tb02495.x [DOI] [PubMed] [Google Scholar]
  • 15. Norman G, Sherbino J, Dore K, et al. The etiology of diagnostic errors: a controlled trial of system 1 versus system 2 reasoning. Acad Med. 2014;89(2):277‐284. doi: 10.1097/ACM.0000000000000105 [DOI] [PubMed] [Google Scholar]
  • 16. Norman GR, Brooks LR. The non‐analytical basis of clinical reasoning. Adv Health Sci Educ Theory Pract. 1997;2(2):173‐184. doi: 10.1023/A:1009784330364 [DOI] [PubMed] [Google Scholar]
  • 17. Custers EJ. Thirty years of illness scripts: theoretical origins and practical applications. Med Teach. 2015;37(5):457‐462. doi: 10.3109/0142159X.2014.956052 [DOI] [PubMed] [Google Scholar]
  • 18. Tichelaar J, Richir MC, Avis HJ, Scholten HJ, Antonini NF, De Vries TP. Do medical students copy the drug treatment choices of their teachers or do they think for themselves? Eur J Clin Pharmacol. 2010;66(4):407‐412. doi: 10.1007/s00228-009-0743-3 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 19. Hensel DTC, Lawless J, Burgess J. A scoping review of Q methodology nursing education studies. Nurse Educ Today. 2022;109:105220. doi: 10.1016/j.nedt.2021.105220 [DOI] [PubMed] [Google Scholar]
  • 20. Churruca KLK, Wu W, Gibbons K, Nguyen HM, Ellis LA, Braithwaite J. A scoping review of Q‐methodology in healthcare research. BMC Med Res Methodol. 2021;21(1):125. doi: 10.1186/s12874-021-01309-7 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 21. Lundberg A dLR, Aliani R. Using Q methodology: sorting out subjectivity in educational research. Educ Res Rev. 2020;31:100361. doi: 10.1016/j.edurev.2020.100361 [DOI] [Google Scholar]
  • 22. Watts S, Stenner P. Doing Q methodological research: theory, method & interpretation. SAGE Publications Ltd; 2012. doi:10.4135/9781446251911 [Google Scholar]
  • 23. EDC C . Castor electronic data capture; 2024. [Available from: https://www.castoredc.com/] [Google Scholar]
  • 24. Banasick S. Kade: a desktop application for Q methodology. J Open Source Softw. 2019;4(36):1360. doi: 10.21105/joss.01360 [DOI] [Google Scholar]
  • 25. Moate RM, Cox JA, Brown SR, West EM. Perceptions of helpfulness of teachers in didactic courses. Counselor Educ Supervision. 2017;56(4):242‐258. doi: 10.1002/ceas.12083 [DOI] [Google Scholar]
  • 26. Brown SR. Political subjectivity, application of Q methodology in political science. Yale University Press; 1980. Available at: https://qmethod.org/1980/01/08/brown-1980-political-subjectivity/ [Google Scholar]
  • 27. Thurstone LL. Multiple factor analysis. Psychol Rev. 1931;38(5):406‐427. doi: 10.1037/h0069792 [DOI] [Google Scholar]
  • 28. Grijpma J, de la Croix A, Meeter M, Kusurkar R. Changes in student appreciation of small‐group active learning: a follow‐up q‐methodological study. Int J Educ Res Open. 2022;3:100199. doi: 10.1016/j.ijedro.2022.100199 [DOI] [Google Scholar]
  • 29. Montori VM, Brito JP, Murad MH. The optimal practice of evidence‐based medicine: incorporating patient preferences in practice guidelines. JAMA. 2013;310(23):2503‐2504. doi: 10.1001/jama.2013.281422 [DOI] [PubMed] [Google Scholar]
  • 30. Grimshaw JM, Thomas RE, MacLennan G, et al. Effectiveness and efficiency of guideline dissemination and implementation strategies. Health Technol Assess. 2004;8(6):iii‐iv, 1‐72. doi: 10.3310/hta8060 [DOI] [PubMed] [Google Scholar]
  • 31. Sinsky CA, Willard‐Grace R, Schutzbank AM, Sinsky TA, Margolius D, Bodenheimer T. In search of joy in practice: a report of 23 high‐functioning primary care practices. Ann Fam Med. 2013;11(3):272‐278. doi: 10.1370/afm.1531 [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 32. Bensing J. Bridging the gap: the separate worlds of evidence‐based medicine and patient‐centered medicine. Patient Educ Couns. 2000;39(1):17‐25. doi: 10.1016/S0738-3991(99)00087-7 [DOI] [PubMed] [Google Scholar]
  • 33. Gabbay J, le May A. Practice‐based evidence for healthcare: clinical mindlines. 1st ed; 2010. [Google Scholar]
  • 34. Ekman I, Swedberg K, Taft C, et al. Person‐centered care‐‐ready for prime time. Eur J Cardiovasc Nurs. 2011;10(4):248‐251. doi: 10.1016/j.ejcnurse.2011.06.008 [DOI] [PubMed] [Google Scholar]
  • 35. Robinson JH, Callister LC, Berry JA, Dearing KA. Patient‐centered care and adherence: definitions and applications to improve outcomes. J Am Acad Nurse Pract. 2008;20(12):600‐607. doi: 10.1111/j.1745-7599.2008.00360.x [DOI] [PubMed] [Google Scholar]
  • 36. Hartjes MG, Richir MC, Cazaubon Y, et al. Enhancing therapeutic reasoning: key insights and recommendations for education in prescribing. BMC Med Educ. 2024;24:1360. doi: 10.1186/s12909-024-06310-4. [DOI] [PMC free article] [PubMed] [Google Scholar]
  • 37. Bakkum MJ, Verdonk P, Thomas EG, et al. A clinical pharmacology and therapeutics Teacher's guide to race‐based medicine, inclusivity, and diversity. Clin Pharmacol Ther. 2023;113(3):600‐606. doi: 10.1002/cpt.2786 [DOI] [PubMed] [Google Scholar]
  • 38. Piët JD, Booth A, Donker EM, et al. Environmentally sustainable prescribing: recommendations for EU pharmaceutical legislation. Lancet Planet Health. e715‐e716. [DOI] [PubMed] [Google Scholar]
  • 39. Danielson S. Q method and surveys: three ways to combine Q and R. Field Methods. 2009;21(3):219‐237. doi: 10.1177/1525822X09332082 [DOI] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Data Availability Statement

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Articles from British Journal of Clinical Pharmacology are provided here courtesy of Wiley

RESOURCES