Physicians’ Use of and Preferences for FDA-Approved Prescribing Information

Abstract

Background:

The Prescribing Information (PI) is the US Food and Drug Administration (FDA)’s primary tool for communicating a summary of the essential scientific information needed for the safe and effective use of a prescription drug to healthcare providers.[1] One challenge with this type of communication is balancing the need to be thorough with the need to be concise.

Objectives:

This study aimed to explore physicians’ preferences for and understanding of specific content and formatting in the PI. This study also explored physicians’ use of and perceptions of the PI.

Methods:

Seventy semi-structured qualitative interviews were conducted with primary care physicians (n = 35) and physicians from a wide range of specialties (n = 35) using web conferencing technology. Using fictitious PI examples, the guide assessed physicians’ interpretation of language and preferences for how certain information is organized and communicated in select sections of the PI. The interview guide also included questions about the resources physicians use to find information about prescription drugs, when and how physicians access the PI, and their perceptions of the PI.

Results:

The findings suggest that of the content and formatting items surveyed, physicians had the greatest preference for: (1) uniformly specifying the age group for which the drug is indicated in the INDICATIONS AND USAGE section, even for medical conditions that are highly associated with only one particular age group (e.g., adult patients), and (2) uniformly including administration information in relation to food (e.g., “with or without food”) in the DOSAGE AND ADMINISTRATION section for drugs with oral dosing. The findings also suggest that including a long list of interacting drug examples in the DRUG INTERACTIONS section may be misinterpreted to be a comprehensive list.

Conclusion:

This qualitative research suggests physicians may prefer more clarity in some sections of the PI.

1. Introduction

The Prescribing Information (PI), a type of FDA-approved labeling, is the US Food and Drug Administration’s primary tool for communicating a summary of the essential scientific information needed for the safe and effective use of a prescription drug to healthcare providers.[1] The PI must be informative and accurate and neither promotional in tone nor false or misleading; and the PI must be updated when new information becomes available that causes the PI to become inaccurate, false, or misleading.¹

There is a tension in this type of drug communication between being thorough (to provide all relevant information) and being concise (to acknowledge the audience’s limited time).^2–3 One commentary suggested that there is a need to clarify the information in the PI for clinicians regarding drug benefits and harms.⁴ Other commentaries have suggested that more can be done to simplify the DRUG INTERACTIONS section of the PI and make it user-friendly for busy prescribers who may not have the time or pharmacological expertise for full comprehension.^5–6 The PI is developed by the pharmaceutical company and reviewed and approved by FDA in accordance with labeling regulations and guidance recommendations. However, since the PI is written for the healthcare provider, it would be useful to understand how the intended audience interprets the PI and if there are physician preferences for how information is communicated in the PI.

In addition, it is unclear how healthcare providers access and use the PI, and if they consider some PI content to be more useful than other content. Available literature offers evidence of prescriber behavior^7–9, prescribing errors^10–12, and prescriber understanding of risk information¹³, whereas only a few studies have directly examined prescriber perceptions of the PI.^14–15 One study noted that the sections in the PI that physicians found to be most useful were those with practical implications for their prescribing decisions like the INDICATIONS AND USAGE, DOSAGE AND ADMINISTRATION, and CONTRAINDICATIONS sections.¹⁴ That study also presented initial evidence that viewing certain sections of the PI was associated with physicians’ drug perceptions and prescribing intentions.

This study explored physicians’ perceptions or interpretations of specific information in five parts of the PI: INDICATIONS AND USAGE, DOSAGE AND ADMINISTRATION, WARNINGS AND PRECAUTIONS, ADVERSE REACTIONS, and DRUG INTERACTIONS. To provide context for the findings on physicians’ interpretation of language in the PI, the current study began by assessing (1) whether the PI is a regular source of drug information for physicians, (2) how and when physicians access and use the PI, and (3) physicians’ perceptions of the PI.

2. Methods

Seventy one-on-one online interviews were conducted with primary care physicians (PCPs) (n = 35) and specialist physicians (n = 35). Individual interviews are valuable when exploring individuals’ opinions, attitudes, and behaviors because they elicit detailed responses that can be difficult to capture in large surveys.^16–18 They also allow more in-depth discussion by each participant than would be possible in a survey or even a focus group.

2.1. Participants

Participants were US physicians (PCP or specialist) who spoke English and prescribed at least 50 drugs per week (including prescription refills). Physicians who had participated in a focus group or interview during the previous three months or who reported working for a market research firm, a pharmaceutical company, the U.S. Department of Health and Human Services, or RTI International (a nonprofit research organization involved in data collection) were excluded. Participants were recruited via phone and email from an existing healthcare panel of physicians who have agreed to be contacted for participation in research studies. Recruitment materials stated that the study was about prescription drug labeling. To ensure a mix of physicians in terms of gender, race/ethnicity, and age, we used soft targets based on recent data sources.^19–21

2.2. Data Collection

Interviews were conducted by RTI International interviewers experienced in qualitative research methods between December 2019 and February 2020 using web conferencing technology. Interviewers received project-specific training prior to data collection. Participants were able to talk with the interviewer and see the interviewer’s screen, which displayed fictitious examples of various sections of the PI. On average, the interviews took 45 minutes to complete.

The interviewer used a semi-structured interview guide to conduct the interviews (see Appendix). We developed the interview guide to address the research questions, with input from medical and regulatory advisors and labeling specialists. The interview guide was identical for PCPs and specialists and included open-ended questions and follow-up probes. All interviews were audio recorded to produce verbatim transcripts. One note-taker documented preliminary themes during the interview. The interview had four parts presented sequentially:

1. resources physicians use to find drug information;

2. when and how physicians access and use the PI;

3. physicians’ broad perceptions of the information in the PI. Participants viewed an example Highlights of Prescribing Information section and a table describing the organization and content of the PI²² and were asked to indicate which PI sections are (a) the most useful, (b) lack information, (c) have too much information, and (d) are confusing based on their experiences in reviewing the PI for drugs they typically prescribe; and

4. physician’s perceptions and interpretation of fictitious examples of specific PI content and formatting developed by FDA labeling subject matter experts. Six fictitious examples are described below.

In the first example (Figure 1), participants were shown three differently worded statements about use of a drug in patients with severe renal impairment that included terminology that may be used in the PI (Figure 1): “contraindicated,” “avoid use,” and “not recommended.” Participants were then asked if the messages conveyed by the three statements with the above terminology were the same or different. After they provided their interpretation, participants were asked if they would consider using the drug (i.e., Drug-X) in a patient with severe renal impairment based on each statement.

Fig. 1.

Contraindications, Avoid Use, and Not Recommended Terminology in the PI

For the second example (Figure 2), participants reviewed two different statements in the INDICATIONS AND USAGE section of the PI for a drug approved to treat a disease in adults that primarily occurs in older adults (i.e., the disease rarely occurs in pediatric patients). The first statement specifically stated that the drug is indicated for the treatment of adults with the disease, whereas the other statement omitted the word “adult.” Participants were asked whether they thought the term “adult” needs to be included and how they would interpret the indication without it.

Fig. 2.

Specifying Age Groups in the INDICATIONS AND USAGE Section

For the third example, participants reviewed two statements in the DOSAGE AND ADMINISTRATION section of the PI related to directions about when the drug should be administered with or without food (Figure 3). The first statement included explicit instructions that the drug was to be taken once daily with or without food. The second statement mentioned only that the medication was to be taken once daily (omitting “with or without food”). Participants were asked to interpret the content and indicate their preferred statement.

Fig. 3.

Including Information in the DOSAGE AND ADMINISTRATION Section Regarding Administration With or Without Food

For the fourth example, participants were asked to review two ways of presenting clinically significant adverse reactions and risk information in the WARNINGS AND PRECAUTIONS section of the PI. In Example A (Figure 4a), the information was ordered by first presenting a description of the clinically significant adverse reaction (labeled paragraph 1 in Example A), then the risk factors (labeled paragraph 2 in Example A), and finally, the steps to take to prevent or manage the adverse reactions (labeled paragraph 3 in Example A). In Example B (Figure 4b), the same information was ordered by first presenting the steps to prevent or manage the adverse reactions (which is paragraph 3 in Example A), the description of the clinically significant adverse reaction (which is paragraph 1 in Example A), and then the risk factors (which is paragraph 2 in Example A). The wording was identical in both examples, but the order of the information was different. Participants were asked to indicate which format they preferred and why.

Fig. 4a.

Warnings and Precautions: Example A

Fig. 4b.

Warnings and Precautions: Example B

For the fifth example, participants reviewed two approaches for listing the most common adverse reactions in the ADVERSE REACTIONS section in the PI (the adverse reactions were identical in both examples, but they were organized differently). In Example A (Figure 5a), the most common adverse reactions associated with the drug were ordered by body system and then by decreasing frequency within each body system. In Example B (Figure 5b), the most common adverse reactions associated with the drug were ordered by decreasing frequency (there was no reference to body systems). Participants were asked to indicate which example they preferred and why.

Fig. 5a.

Adverse Reactions: Example A

Fig. 5b.

Adverse Reactions: Example B

For the sixth example, participants viewed four different approaches of incorporating drug interaction information into the DRUG INTERACTIONS section (Figure 6). The first sentence, which included the category of interacting drugs, was identical across the four different approaches (Examples A-D). Example A did not include any additional information. Examples B, C, and D included an additional sentence that provided examples of interacting drugs or a method of obtaining more information about example interacting drugs. The additional sentence in Example B included an abbreviated list of representative examples of the interacting drugs. The additional sentence in Example C included a more detailed list of representative examples of interacting drugs. Finally, the additional sentence in Example D did not include examples of specific interacting drugs but instead included a hyperlink to a fictitious FDA drug interactions webpage containing examples of interacting drugs. Participants were asked to indicate which of the four examples they preferred and why.

Fig. 6.

Providing Examples of Interacting Drugs in the DRUG INTERACTIONS Section

2.3. Data Analysis

Qualitative analysis software (NVivo version 11; QSR International) was used to organize and code the verbatim transcripts. Based on the interview guide and initial interview themes, a codebook was developed to use in data analysis, which was refined during the coding process. The development of codes fell into one of two categories: (1) predetermined codes (e.g., codes for questions that asked participants to select a preference out of two or more possible options) and (2) emergent codes, for which the coding was developed after an initial review of the data based on participants’ responses (e.g., codes for questions that asked participants for their interpretation of specific information in the labeling). To start, seven transcripts were randomly selected for two team members to review and code independently. The two coders then met with two other project team members to identify and resolve discrepancies in the coding of these transcripts. After achieving an average Cohen’s kappa of 0.95 (range: 0.93–0.98), the two coders split the remaining transcripts for coding and met regularly to discuss any emergent codes.

The project team members identified themes in participant responses and looked for differences between PCPs and specialists; the next section notes the few differences between the two groups. In line with previous qualitative research, the next section uses the following qualitative terms to indicate the extent to which findings were observed across participants: none, a few, some, approximately one-half, many, and most.^23–24

3. Results

3.1. Participant Characteristics

Interviews were conducted with 70 physicians: 35 PCPs (including family practice, internal medicine, or general practitioner) and 35 specialists from 12 different medical specialties: obstetrics and gynecology (n = 5), pediatrics (n = 4), psychiatry (n = 4), critical care (n = 1), emergency medicine (n = 2), oncology/hematology (n = 3), neurology (n = 3), rheumatology (n = 3), orthopedics or sports medicine (n = 3), gastroenterology (n = 2), allergy/immunology or infectious disease (n = 2), dermatology (n = 2), and endocrinology (n = 1). About half of the participants (53%) were male, 54% were White, 23% were Asian, 10% were Black, and most were non-Hispanic (91%). Most participants had practiced medicine for over 11 years and spent between 90%–100% of their time in direct patient care (Table 1).

Table 1.

Participant Characteristics: Primary Care Physicians (PCPs) and Specialists

	Total (n = 70)		Provider Type
			PCPs (n = 35)		Specialists (n = 35)
	N	%	N	%	N	%
Percentage Time in Patient Care
50–74	3	4.3	3	8.6	0	0.0
75–89	5	7.1	2	5.7	3	8.6
90–100	62	88.6	30	85.7	32	91.4
Years in Practice
Less than 5 years	7	10.0	4	11.4	3	8.6
5–10 years	9	12.9	4	11.4	5	14.3
11–20 years	21	30.0	8	22.9	13	37.1
21–30 years	21	30.0	11	31.4	10	28.6
31 or more years	12	17.1	8	22.9	4	11.4
Practice Size
Solo	13	18.6	10	28.6	3	8.6
Small group practice	30	42.9	11	31.4	19	54.3
Large group practice	27	38.6	14	40.0	13	37.1
Practice at Academic Hospital
Yes	15	21.4	6	17.1	9	25.7
No	55	78.6	29	82.9	26	74.3
Sex
Female	33	47.1	18	51.4	15	42.9
Male	37	52.9	17	48.6	20	57.1
Hispanic/Latino
Yes	6	8.6	5	14.3	1	2.9
No	64	91.4	30	85.7	34	97.1
Race
White	38	54.3	15	42.9	23	65.7
Black/African American	7	10.0	3	8.6	4	11.4
Asian	16	22.9	11	31.4	5	14.3
Other	7	10.0	5	14.3	2	5.7
Prefer not to say	2	2.9	1	2.9	1	50.0
Region
Northeast	12	17.1	5	14.3	7	20.0
Midwest	14	20.0	8	22.9	6	17.1
West	10	14.3	7	20.0	3	8.6
South	34	48.6	15	42.9	19	54.3

3.2. Resources Physicians Use to Find Drug Information

When asked about the resources they used to find drug information, most participants mentioned using electronic resources including UpToDate, Epocrates, Physicians’ Desk Reference (PDR), the manufacturer websites, other health-related websites (e.g., Drugs.com, WebMD, PsycNET), government sites (e.g., National Institutes of Health) and online journals. Participants also made references to using smart-phone applications for some of these resources. One PCP remarked:

“I use Epocrates and MPR [Monthly Prescribing Reference], so they’re in my phone. I use them every day for products that I’ve used in the past that sometimes I forget the dosing or if I need to make any renal adjustment or are they safe for pregnant women, that kind of stuff, or stock dose.”

(PCP, 59, Female)

Some participants mentioned using resources embedded in their electronic health records (EHR; e.g., Epocrates, UpToDate, Medicom and Micromedex). Fewer participants mentioned paper resources such as reference books like the Tarascon Pocket Pharmacopoeia and the Red Book. Only about a fifth of participants specifically mentioned using the PI (both the online version and the paper version) to review information about drugs they have prescribed before. Note, however, that the PI is reprinted in the PDR or excerpted in other electronic references that participants reported using.

When participants were asked if the information and resources they use differ if they are looking up information about a drug they have not prescribed before, participants often said that they would start with the same resources, but they would perform a more thorough review of the resources for a drug they had not prescribed before. Some participants said they would supplement their usual resources by reviewing additional resources, specifically mentioning the paper version of the PI but also journal articles and talking to colleagues or pharmaceutical representatives. Participants also noted that sometimes the resources they use depend on the type of drug information they are looking for (e.g., information on drug pricing is not included in all resources).

3.3. When and How Physicians Access the PI

All participants reported previous experience with the PI. In many cases, participants said they would refer to a drug’s PI when prescribing a drug they have never prescribed before or when prescribing a drug they do not typically prescribe. One PCP explained:

“I definitely look at that [PI] when I am not familiar with the drug. So, I have to read everything about the drug as an introduction.”

(PCP, 60, Female)

In instances when they are familiar with the drug, some participants said they may look at the drug’s PI as a refresher or to confirm their understanding of the indications, dosing, adverse reactions, drug interactions, or contraindications. One specialist commented:

“It’s to verify information that I’m already aware of, or to seek contraindications for certain medical conditions, or medications… that can’t be used with that medication, or there are certain risk associated with, say diabetes, and that medication or something like that.”

(Orthopedic Surgeon, 55, Male)

When asked how they access the PI, most participants said they access it online by going directly to the manufacturers’ website or by searching on Google for the drug name. However, some participants noted that they will sometimes use the paper PI. A few participants stated that they use the PI when interacting with patients to provide a rationale for prescribing the drug, to address specific concerns about a drug, or to assess the likelihood that patient-reported side effects are caused by the drug. In addition, a few participants have found referring to the paper PI to be faster and easier than other methods, like searching online. One PCP explained:

“So, even with drugs I’m using, sometimes if I have a patient complaint, I’ll actually pull out the [PI] from a sample that I have. It’s a little quicker than going online and easier showing it to them; and then they can even take it home with them and think about it more.”

(PCP, missing age, Male)

3.4. Physicians’ Perceptions of the Information in the PI

When asked which section or sections of the PI were most useful in their practice, most participants said DOSAGE AND ADMINISTRATION (especially for a less-familiar drug). About one-half of the participants cited DRUG INTERACTIONS (particularly when patients are taking other prescription drugs), INDICATIONS AND USAGE (particularly to make sure the drug is indicated for the patient’s condition), CONTRAINDICATIONS (particularly to ensure they do not harm the patient by prescribing a drug the patient should not take), and ADVERSE REACTIONS (also to ensure that the drug doesn’t harm the patient) as being most useful. One PCP remarked:

“This almost kind of seems to go in order of importance. Dosing administration, … contraindications, adverse drug interaction… Because I think those are the most important things that we would look for. It kind of gives you the information of what the drug is, what’s the dosing, what are the indications, how’s it administered, major things to keep in mind in terms of contraindications of the drug or reactions or interactions.”

(PCP, 38, Female)

Most participants said there are no sections that do not provide enough information. They noted that the PI is thorough and provides the information they need to make prescribing decisions. Some participants cited a few sections as potentially having too much information, including ADVERSE REACTIONS, CLINICAL PHARMACOLOGY, and, NONCLINICAL TOXICOLOGY. However, most participants stated that even though some sections of the PI might contain information they do not use, including this information in the PI is important, and they can skim information that they consider less relevant.

When asked which sections of the PI include confusing information, most participants said they did not find any of the information in the PI confusing; however, a few stated that the CLINICAL PHARMACOLOGY and NONCLINICAL TOXICOLOGY sections can be difficult to understand because of their technical nature. One PCP commented about the PI:

“No, I mean I think this has been pretty well honed and refined through the years and is pretty standard. I wouldn’t … Since it’s what we’re used to seeing I wouldn’t monkey with it.”

(PCP, 63, Male)

3.5. Examining Specific PI Language and Formatting/Organization

3.5.1. Interpreting “contraindications”, “avoid use” and “not recommended” terminology in the PI (Figure 1).

Most participants differentiated between the three statements used, with Statement 1, which included the term “contraindicated,” being the strongest warning not to use Drug-X in patients with severe renal impairment. Participants considered Statement 2, which included the term “avoid use,” as a step down from Statement 1, and Statement 3, which included the term “not recommended,” as a less absolute warning to using the terminology “avoid use.” One specialist remarked:

“Number one, Drug-X is contraindicated. It means that basically, that the patient with severe renal impairment should not be getting that medication at all. Number two, the avoidance, avoid, meaning that even though there’s severe renal impairment, if maybe all other treatment alternatives have been exhausted or the patient, that’s the only medication that’s available for treatment of that particular condition, then it might be possible to defend the use of the medication […], number three there, is sort of the least rigorous of the three. It’s not recommended, but it doesn’t mean that we absolutely shouldn’t use it, and that it probably could be used if the benefit risk ratio was significantly on the benefits side and obviously would need … warning of the patient and extensive monitoring to make sure that levels don’t rise to dangerous levels because of the renal impairment.”

(Rheumatologist, 58, Male)

Consistent with their interpretation of the terminology, most said they would not consider using Drug-X in a patient with severe renal impairment based on Statement 1. Half of all participants said they would consider using Drug-X in patients with severe renal impairment based on Statement 2. These participants said they may be hesitant to use Drug-X because the statement says to “avoid use in patients with severe renal impairment,” but it would be a possibility depending on other circumstances. About two-thirds of participants said they would consider using Drug-X based on Statement 3 because the statement does not directly say that it is contraindicated or that it should be avoided.

3.5.2. Specifying the age groups in the INDICATIONS AND USAGE section (Figure 2).

The majority of participants said that the INDICATIONS AND USAGE section for a drug indicated for adults needs to include the term “adults.” More specialists than PCPs said they thought that the PI needs to explicitly include the term “adults.” Participants noted that it would be clearer to include the term “adults” in the indication even in cases where the drug treats a condition that does not or rarely occurs in pediatric patients. One specialist remarked:

“Multiple myeloma is usually in an aged population. But, having said that, there’s always exceptions […]. So I like the fact of example 1 a lot better than example 2 because it’s more specific. And I think the more specificity you add to product information …, the better off patients, physicians, and their families will be.”

(Dermatology, 62, Male)

When asked how they would interpret the indication if it did not specifically reference adults, some participants (about a third) said they would assume that the drug could be used in any age group, some participants (also about a third) said they would assume that the medication should only be used in adults, and the rest of the participants said they were unsure how to interpret the information and would need to do more research or did not provide a specific answer to this question.

3.5.3. Instructions in the DOSAGE AND ADMINISTRATION section about when the drug should be administered with or without food (Figure 3).

When asked how they would interpret the instructions that did not include the phrase “with or without food” most participants said they would interpret those instructions to mean that it did not matter if patients took the drug with or without food. However, most participants said that the PI should explicitly state that a drug can be taken with or without food in a prominent place in the labeling (if this was accurate) as shown in the first statement. One PCP commented:

“It’s always good. Yes, because again, it has a clinical impact and number two, it also answers my patient’s question. It’s a very common question amongst my patients.”

(PCP, 54, Male)

These participants explained that instructions should be clear and straightforward to avoid confusion and noted that their patients frequently ask whether they need to take drugs with food, so having explicit instructions in the PI would save physicians time because they would not need to use other resources to find this information. More PCPs than specialists preferred the inclusion of explicit instructions about taking the drug with or without food in this section.

3.5.4. Order of Information in WARNINGS AND PRECAUTIONS (Figures 4a and 4b).

Slightly more participants preferred Example B over Example A, citing that they liked to see risk management information before prescribing the drug. When risk management information is presented last, some participants thought that they might stop reading and miss that information altogether. One specialist remarked:

“I think the second one gets to the nuts and bolts first. ‘Before you do this, consider this.’ That’s the first reason to pause, hesitate, before prescribing something right away. Then the explanation, a little bit more deeper.”

(Allergy/Immunology, 44, Female)

Participants who preferred Example A said they liked it because it followed a logical order, beginning with a description of the warning(s), then the consequences, and then what to do about the consequences. This sequence resonated with how these participants think about the risks. As one participant noted, Example A presents the “why” first and then tells the provider what to do next, whereas Example B provides directions without first presenting the “why.” One specialist remarked:

“I’d like to know, from the beginning, what I’m being concerned about. … And then, I’d like to know, once I’ve been convinced that this drug is a problem with blood pressure, then I want to get into the recommendation on how to approach the drug.”

(Oncology, 61, Male)

3.5.5. Organizing adverse reactions in the ADVERSE REACTIONS Section (Figures 5a and 5b).

Participants’ preferences were split between the two approaches of organizing adverse reactions. About one-half of participants preferred Example B (adverse reactions listed by decreasing frequency without ordering by body system too). These participants said they preferred having the most common adverse reactions at the top of the list and the least common at the bottom. Participants said it was helpful to know the most common adverse reactions to consider in a clinical situation and for summarizing potential adverse reactions for patients. These participants indicated they put more importance on how often the adverse reactions occur rather than the body systems impacted. One PCP explained:

“I prefer example screen B … because the most common adverse events … [are] listed on top. So I would tell the patient if I start this new drug: ‘Oh main thing you’ll maybe get headaches, sinusitis, abdominal pain and fever, and candida infection is rare.’ ……the other one is confusing, takes more time to read, go down everything.”

(PCP, 58, Female)

Close to half of participants preferred Example A (i.e., listing adverse reactions by body systems and decreasing frequency for each body system) and noted that the layout and organization of the table, including bolding and indentation of symptoms, made the table easier to read than a “laundry list” of adverse reactions. Participants who preferred this option also noted that Example A provides a conceptual organization for adverse reactions, which resonates with how doctors are trained to look at symptoms by body system. Some participants noted that this format is useful when looking for a specific symptom, particularly when patients call with symptoms and use different terms to describe them. A table organized by body system helps orient the provider to the affected body system. One specialist remarked:

“I like “A” better because it categorizes it according to symptoms. So if I’m looking for something, particularly a reaction, I can hone in right on where it is. I have to go searching through it in “B”. It is not symptom specific.”

(Endocrinology, 55, Male)

3.5.6. Presenting information in the DRUG INTERACTIONS SECTION (Figure 6).

The majority of participants preferred Example C, which included a more detailed list of representative drug interaction examples. Among those who preferred Example C, the majority misinterpreted the information as all-inclusive and therefore not requiring additional time to review other resources for more complete information on interacting drugs (in fact, the list of interacting drugs provided in Example C was not comprehensive). They noted that it is helpful to have CYP3A inhibitors listed and to avoid having to guess or recall from memory. One specialist remarked:

“This is really clear in my mind. I really like Example C, because although all of us know mode of actions of a lot of drugs, but we may not be familiar with other substances that could increase the side effects or adverse events in drugs…I think it’s important, especially when you’re talking about drug-drug interactions, to be as complete and thorough as is possible because drug-drug interactions occur not only with other drugs, but can occur with things we don’t think of as a drug, like foods. So, I think that to be thorough and complete for the patient’s benefit, for the patient’s family’s benefit as well as the physician’s benefit, I think thoroughness is the watchword and the key.”

(Dermatology, 62, Male)

A few participants preferred Example D, which included a hyperlink to a fictitious FDA drug interactions website containing a list of interacting drugs. These participants preferred this example because it provides a brief, concise summary of the key points but also gives an option to see all available information. They also noted that the hyperlink is helpful because the website could potentially provide more complete and up-to-date information. One PCP remarked:

“Format D might be the best. Short, direct to the point, and if you want to explore further, you have a hyperlink there that you could click on …”

(PCP, 54, Male)

4. Discussion

This study used in-depth interviews to investigate how physicians use and perceive the FDA-approved PI for prescription drugs as well as physicians’ preferences and understanding for specific PI content and format. This study yielded several noteworthy findings regarding these topics.

First, consistent with findings from past studies, this study found that physicians used a variety of sources to look up drug information.^25–28 Notably, the PI was not listed as a top source of information; instead, participants reported most often using Epocrates and UpToDate. Our findings suggest that the PI serves more as a supplementary source of information rather than the primary one. However, it is likely that the PI is actually used more frequently than what this study found as physicians may be viewing information from the PI through other sources that are based on the PI.

Second, despite not being mentioned as a top information source, when asked about the PI content, participants reported that the PI works reasonably well in meeting their needs as far as the amount of information and ease of understanding. Future research may explore the reasons why physicians may not be using the PI as a primary source of drug information despite having positive perceptions of the information included.

When participants were asked about their perceptions and interpretations of different language in fictitious examples, the findings suggested two scenarios that may affect physicians’ interpretation. First, when the INDICATIONS AND USAGE section does not specify the age groups (e.g., whether the drug is indicated in adults), some participants assumed that the drug could be used in any age group and a few participants said they would have to do additional research to determine the population for which the drug is indicated. This finding suggests that adding the indicated age groups to indications would help clarify the approved age groups for physicians. In addition, researchers could examine other assumptions that physicians hold and the heuristics at play when processing information in the PI.

Second, when comparing the lists of interacting drugs, most participants appeared to assume the longest list was comprehensive, which it was not. This raises the question as to whether physicians assume that when many example interacting drugs are included in the DRUG INTERACTIONS section, the list is comprehensive. Research could explore whether there is a threshold at which physicians are more likely to assume a list is comprehensive and how those assumptions impact prescribing decisions. Research could also explore whether additional language in the PI could effectively alert physicians to the limited nature of the list.

Most participants interpreted “contraindicated,” “avoid use,” and “not recommended” terminology consistent with FDA regulations and guidances.^29–30 This finding adds to the wider literature on efforts to ensure drug information promotes accurate prescribing and medication safety.^31–33 The feedback on information layout and organization of the WARNINGS AND PRECAUTIONS section and the most common adverse reactions table in the ADVERSE REACTIONS section was mixed, with participants offering compelling reasons for preferring information presented in different orders. It may be that physicians prefer different formats based on clinical context. In the current technology landscape, electronic formats have become more prevalent and offer opportunities for users to choose their preferred format from a list of options. Future research could examine this in more detail to further elucidate physician preferences for drug information formats.

Finally, there are competing demands on physician time. Unsurprisingly, a theme that emerged throughout the various parts in the interview was participants’ preference for complete and accurate information and formats that would save time. Participants appreciated when they did not have to spend additional time deciphering unclear instructions or trying to find other resources for complete drug information. For instance, participants preferred language specifically noting whether the drug is prescribed for adults even in scenarios where the medical condition that it treats rarely occurs among pediatric populations; similarly, participants preferred instructions for whether a drug should be administered with or without food even though most interpreted the absence of instructions to mean that the drug can be taken with or without food. Additionally, participants preferred to have a comprehensive list of interacting drugs rather than no examples or an abbreviated list of examples; although, the fictitious example did not include a comprehensive list of interacting drugs. In all, the findings from this study resonate with past suggestions that physicians may prefer more clarity in some sections of the PI.⁴

4.1. Limitations

As with most qualitative research, these findings are constrained in terms of generalizability. The physicians who participated may not be representative of all physicians. Social-desirability bias may have affected participants’ responses in the interviews. We encourage more quantitative research in this area, such as healthcare provider surveys and experiments with different formats. Another consideration is that the study only included physicians in this sample. Future research could examine whether these results extend to other types of healthcare providers (e.g., physician assistants, nurse practitioners, and pharmacists).³⁴ We did not vary the order of the questions or examples, so it is possible that the order affected participants’ responses. Preferences about the content and wording of other PI sections and the types of information conveyed could also be examined to extend these findings.

4.2. Conclusion

This study offers new insights regarding physicians’ perceptions of the FDA-approved PI. These findings suggest that physicians may prefer the addition of clarifying language in two sections of the PI: (1) specifying the approved age groups in the INDICATIONS AND USAGE section even for medical conditions or diseases that are known to primarily occur in a specific age subgroup, and (2) explicitly stating “with or without food” in the DOSAGE AND ADMINISTRATION section for drugs used orally that can be administered with or without food. These findings also indicate that a long list of examples in the DRUG INTERACTIONS section may be misinterpreted to be a comprehensive list. These insights warrant further investigation to aid in the communication of drug information to healthcare providers.

Appendix

Physician Interviews on FDA-Approved Prescribing Information: Moderator Questions and Examples

Resources to Find Information About Prescription Drugs

1. What kind of information do you typically look up for a drug that you have prescribed before?

   • IF NEEDED, PROBE:

     • Dose?
     • Indication?
     • Contraindications?
2. What resources do you use to look for information about prescription drugs that you have prescribed before?

   • IF NEEDED, PROBE:

     • App on a smart phone
     • Paper in drug package
     • Database within the EHR (e.g. Micromedex)
     • Health-related websites (e.g. Drugs.com)
     • Paper-based book/resource guide
     • Manufacturer’s website

3. Do the resources that you use differ depending on the type of information you are looking for about the drug? If so, how do they differ?

4. Does the kind of information that you look for about prescription drugs differ if you are looking for information about a drug that you have not prescribed before? If so, how?

   1. Do the resources that you use to look for this information differ from the ones you said you use to look for information about drugs that you have prescribed before? If so, how?

5. What resources do you typically use for overdose information?

Accessing and Using Prescribing Information

Next, I’m going to ask you some questions about FDA-approved Prescribing Information, also known as the package insert, drug label or labeling. This is a summary of the information needed to use a prescription drug safely and effectively and is drafted by the drug company and approved by the FDA. Moving forward I’ll refer to the FDA-approved Prescribing Information as the “label.”

• 1Under what circumstances do you look at the drug’s label?
• 2Where do you look for the drug’s label? (IF NEEDED: For example, an online search, Drugs.com, a drug manufacturer’s website, FDA website, DailyMed, paper label in the drug packaging)
• 3How do you typically find out that there has been an update to the label—for example, newly approved indications, new black box warnings or contraindications, new dosages, new risk information, new information about patients with renal impairment, or new drug interactions? (IF NEEDED: “Dear prescriber” letters from drug companies, EHR, pharmaceutical representatives, apps, the label, Drugs@FDA, ads in medical journals, DTC ads, prescriber guides (e.g. Medical Letter, Monthly Prescribing Reference))

Perceptions of Information in the Prescribing Information

The label has three parts:

• Highlights of Prescribing Information (see an example below)

• Table of Contents

• Full Prescribing Information (see a description of the contents of each of the sections in the Full Prescribing Information below)

We’re going to talk about the Full Prescribing Information. Please review and let me know when you’re ready to discuss.

Highlights of Prescribing Information (PLR-Format Labeling)

[Show: Full Prescribing Information Sections of PLR-Formatted Labeling on screen [Table 1 from Gassman, Nguyen, Joffe, 2017]]

• 4What sections or parts of the label do you find most useful in your practice? Why?
• 5
  What sections or parts of the label do not provide enough information?

  1. What information should be added?
• 6
  What sections or parts of the label includes too much information?

  1. What information should be deleted?
• 7Are there sections or parts of the label that are confusing? If so, what should be done to make them clearer?

Interpretation of Language and Formatting in the Prescribing Information

Now I’m going to show you some specific examples of language from the label. After each set, I’ll ask you some follow-up questions. Please read the following three examples and let me know when you’ve finished reading.

EXAMPLE SCREEN

1. Looking at these three statements, are the messages the same or different? Please explain.

2. Would you consider using DRUG-X in a patient with severe renal impairment based on Statement 1? Statement 2? Statement 3? Why or why not?

FDA generally recommends including the age group in the INDICATIONS AND USAGE section.

For this example, DRUG-X is approved to treat certain patients with multiple myeloma. As you know multiple myeloma primarily occurs in older adults.

Please take a moment and read the examples I’m showing on the screen and let me know when you’re ready to discuss.

EXAMPLE SCREEN

1. If a drug is indicated for adults, do you think it needs to explicitly include the term “adults” in the INDICATIONS AND USAGE section – as in Example 1? Why or why not?

2. If the INDICATIONS AND USAGE section did not specifically include “adults” for diseases that typically do not occur or rarely occur in pediatric patients (as in Example 2) how would you interpret this information?

[IF NOT CLEAR FROM Q1 and Q2]. Looking at both examples (Example 1 specifically referencing adults and Example 2 not), which one do you think is best? Why do you say that?

Next, I have a few questions about the best way to include dosage information in the label.

Some labels include specific instructions in the DOSAGE AND ADMINISTRATION section about whether the drug should be administered with or without food. Other labels do not include specific instructions about whether the drug should be administered with or without food Please take a minute and review the examples on the screen and let me know when you’re ready to discuss.

EXAMPLE SCREEN

1. If the DOSAGE AND ADMINISTRATION section said “Take Drug-X orally once daily” – Example b (and did not include “with or without food”) how would you interpret this information?

   1. Would you think it needed to be taken either with food or without food, or would you think it didn’t matter if not specifically stated?

2. If a drug can be taken with or without food, do you think it needs to explicitly state so in the DOSAGE AND ADMINISTRATION section – as in example a? Why or why not?

Information about clinically significant adverse reactions and risks can be presented in different ways in the WARNINGS AND PRECAUTIONS section. Please review the two examples on the screen.

Example a orders information in the following way: (1) description of the clinically significant adverse reaction (highlighted in yellow), (2) risk factors (highlighted in green), and (3) steps to take to prevent or manage the adverse reactions (highlighted in blue).

Example b orders information in the following way: (1) steps to take to prevent or manage the adverse reactions (highlighted in blue) (2) description of the clinically significant adverse reaction (highlighted in yellow), (3) risk factors (highlighted in green).

Note that the wording of each section is identical in both examples.

Please let me know when you’ve finished reading.

EXAMPLE SCREEN

1. Looking at the information in the WARNINGS AND PRECAUTIONS section in Examples a and b, which format do you prefer and why?

Now I’d like to ask you some questions about drug risk information included in the label. The ADVERSE REACTIONS section contains a listing of the adverse reactions associated with the drug or related drugs. Adverse reactions can be listed in the ADVERSE REACTIONS section in different ways.

Note the two examples contain the identical adverse reactions but they are presented in two different ways:

1. In Table a, the most common adverse reactions associated with Drug-X are ordered by body system and then by decreasing frequency (example #1)

2. In Table b, the most common adverse reactions associated with Drug-X are ordered by decreasing frequency (example #2)

3. Please let me know when you’ve finished reading.

EXAMPLE SCREEN

1. Looking at how the adverse reactions of Drug-X are presented in the ADVERSE REACTIONS section of the label in Table a and b, which version do you prefer and why?

The DRUG INTERACTIONS section of the label should include clinically significant known and potential drug interactions, mechanisms of the drug interactions, clinical implications of the drug interactions, and practical instructions to prevent and manage the drug interactions.

Please take a look at the following examples showing different ways of incorporating this information in the label. Note that the examples are identical except for the additional highlighted sentence at the end of each. Please let me know when you’ve finished reading.

EXAMPLE SCREEN

1. Which format do you like best and why?