Interviews and Focus Group Research

Introduction

Interviews and focus groups are among the most widely used methods for generating qualitative evidence in health research. They enable researchers to examine how patients, families, clinicians, managers and policy actors understand health, illness, treatment, service quality and organisational change. Unlike surveys, which generally predetermine response categories, qualitative inquiries allow participants to explain meanings, experiences and interpretations in their own terms.

For health-care managers and researchers, these methods are especially valuable when investigating questions such as: Why do patients discontinue treatment? How do staff experience a new electronic health record? Why do communities distrust a vaccination programme? How do informal practices influence patient safety? The quality of the findings depends not only on asking questions, but also on sampling, rapport, probing, facilitation, reflexivity, ethics, recording, remote practice and systematic analysis.

An interview is a social encounter in which knowledge is produced through interaction between researcher and participant. A focus group is a moderated discussion in which data arise both from individual contributions and from group interaction. Neither method is automatically superior. The appropriate choice depends on the research question, population, sensitivity of the topic, desired depth and practical setting.

Learning objectives

By the end of this chapter, the learner should be able to:

1. Distinguish structured, semi-structured, unstructured, in-depth and key-informant interviews.
2. Explain the distinctive logic, strengths and limitations of focus groups.
3. Design a coherent interview or discussion guide aligned with research objectives.
4. Apply effective probing, listening, rapport management and facilitation techniques.
5. Plan ethical, inclusive and secure face-to-face and remote data collection.
6. Critically evaluate data quality, researcher reflexivity and saturation claims.
7. Apply these methods to health-care management and service-improvement problems.

1. Conceptual foundations

1.1 Interviews as co-produced knowledge

Qualitative interviews do not provide a transparent recording of an objective reality. Participants interpret their experiences, and researchers interpret participants’ accounts. Meaning is therefore co-produced within a social, cultural and institutional context. A patient may describe a hospital differently when speaking to a university researcher, a nurse or a hospital manager. Similarly, an employee may avoid criticism if the interviewer is perceived to represent senior management.

This does not make interviews unreliable. It means that the conditions of knowledge production must be documented and critically examined. The researcher should consider who asked the questions, how the relationship affected the interaction, what was not said and how the account relates to observed practices or documentary evidence.

1.2 Interviews and focus groups compared

The choice should be theoretically and ethically justified rather than based only on convenience. A focus group may reveal that patients negotiate the meaning of “good care” collectively, whereas individual interviews may reveal experiences of stigma that participants would not disclose publicly.

2. Types of interviews

Structured interviews

A structured interview uses a fixed sequence and wording, with limited deviation. It improves consistency across interviewers and can be useful when comparing multiple facilities or obtaining clearly specified information. However, it restricts unexpected exploration and is closer to an orally administered questionnaire than to an interpretive interview.

Semi-structured interviews

Semi-structured interviews use an interview guide containing topics and suggested questions, while allowing the interviewer to change the order, ask follow-up questions and explore unanticipated issues. This is often the most useful format for health-services research because it combines comparability with flexibility.

In-depth interviews

An in-depth interview is designed to obtain detailed accounts of experience, meaning and decision-making. It may be semi-structured or relatively open-ended. In-depth work requires time, active listening and tolerance of pauses. The interviewer should not confuse depth with length: a focused 45-minute interview may be more valuable than a poorly managed two-hour conversation.

Key-informant interviews

Key informants possess specialised knowledge because of their role, experience, social position or involvement in a programme. Examples include a district health officer, infection-control lead, traditional birth attendant, patient advocate or laboratory manager. Key-informant interviews are useful for understanding systems, policies and implementation processes, but they may privilege official or elite perspectives. Their accounts should not be treated as representative of an entire population without corroboration.

Narrative and life-history interviews

Narrative interviews invite participants to construct an extended account of events over time. They are appropriate for chronic illness trajectories, professional careers and experiences of long-term care. The interviewer supports the narrative without imposing a rigid chronology, then may ask clarifying questions.

3. Focus group research

A focus group is a planned discussion involving a small number of participants who share relevant characteristics or experiences. The moderator introduces topics, encourages interaction and protects the conditions for respectful participation. The group is not simply several interviews conducted simultaneously. Agreement, disagreement, silence, humour, hesitation and the way participants respond to one another are all potential data.

Appropriate uses

Focus groups are particularly useful when the research concerns:

• Shared language and community concepts of illness.
• Reactions to a proposed service, policy or patient-information material.
• Norms surrounding health-seeking behaviour.
• Differences between professional groups or user groups.
• Explanations for survey findings.

They are less appropriate for highly sensitive disclosures, experiences involving violence, identifiable complaints against colleagues or topics where participants have substantial power differences.

Composition and size

Groups should be sufficiently homogeneous to support discussion but sufficiently varied to generate meaningful perspectives. A group of senior physicians and junior nurses may inhibit discussion because of hierarchy. Separate groups may therefore be appropriate. Homogeneity should be considered in relation to age, gender, professional role, language, socioeconomic position and experience of the service.

A practical group often contains approximately five to eight participants. Smaller groups may be suitable for complex or specialised topics; larger groups are harder to facilitate and may reduce participation. The number of groups is determined by the diversity of perspectives and the point at which additional groups contribute little new conceptual insight, not by a universal numerical rule.

The moderator and assistant

The moderator guides the discussion without becoming its main speaker. An assistant may observe group dynamics, manage recording and take field notes. The assistant should not distract participants or create the impression of surveillance. At the conclusion, the research team should record contextual observations, including seating, interruptions, emotional shifts and interactions between participants.

4. Designing an interview or focus-group guide

A guide is a flexible research instrument, not a script that must be read mechanically. Its design should proceed from the research question to domains, questions and probes.

Step 1: Clarify the inquiry

Begin with a precise question. “What do patients think about the hospital?” is too broad. “How do adults with diabetes experience continuity of care after discharge?” identifies a population, phenomenon and context. The question should indicate whether the study seeks experiences, meanings, processes, explanations, expectations or evaluations.

Step 2: Identify domains

Domains are major areas of inquiry derived from the objectives or conceptual framework. For a study of telemedicine, domains might include access, usability, privacy, clinical communication, trust, cost and future preferences.

Step 3: Write open, comprehensible questions

Strong questions are open-ended, neutral and singular. Compare:

• Weak: “Did the staff explain the medication properly?”
• Better: “What did the staff explain about your medication?”
• Follow-up: “What, if anything, remained unclear?”

Avoid jargon, double-barrelled questions, presuppositions and questions that imply a desirable answer. “How satisfied were you with the excellent new service?” contains an assumption and a leading evaluation.

Step 4: Organise the sequence

A useful sequence is:

1. Welcome, consent and practical explanation.
2. Easy contextual questions.
3. Core experience questions.
4. Clarification and probing.
5. Reflective or evaluative questions.
6. Sensitive issues, when rapport has developed.
7. Summary, correction and closing.

The “funnel” approach moves from broad questions to more focused questions. However, the interviewer should follow the participant’s account when doing so produces a richer understanding.

Step 5: Add probes

Probes deepen or clarify an answer without supplying content. Common probes include:

• “Can you tell me more about that?”
• “What happened next?”
• “Can you give an example?”
• “How did that affect you?”
• “What did you mean by ‘difficult’?”
• “Were there other views in your team?”
• “What would have made the experience different?”

A probe should be purposeful. Repeatedly saying “why?” can sound accusatory; “What led to that decision?” is often less judgmental.

Step 6: Pilot and revise

Pilot the guide with people similar to the intended participants, or with experienced colleagues where participant piloting is impossible. Assess comprehension, order, emotional burden, duration, translation and whether the questions generate evidence relevant to the objectives. A pilot may reveal that a question is too abstract, culturally inappropriate or interpreted differently than intended.

5. Conducting high-quality interviews

Rapport and role management

Rapport is a working relationship based on respect, clarity and psychological safety. It is not an attempt to become a participant’s friend or to promise outcomes the researcher cannot deliver. At the beginning, explain the purpose, voluntary nature, recording arrangements, confidentiality limits, expected duration and right to stop.

Researchers should manage role expectations carefully. A participant may assume that a health-care researcher can provide treatment, resolve a complaint or influence employment decisions. These assumptions should be corrected respectfully. In organisational research, the interviewer should explain whether managers will receive identifiable comments and how reporting will protect confidentiality.

Listening and silence

Effective interviewing involves more listening than speaking. Attend to words, pauses, emotion, contradictions and changes in tone. Brief silence often gives participants time to formulate a response. Interruptions should be limited, but clarification may be necessary when an account is ambiguous or time is constrained.

Probing without contamination

A good probe expands the participant’s account. A poor probe introduces the researcher’s preferred explanation. For example, “Was the delay caused by staff shortages?” may impose a hypothesis. “What contributed to the delay?” leaves the participant free to identify staffing, transport, communication or other causes.

When an account appears inconsistent, avoid confrontation. Use temporal clarification: “Earlier you mentioned that the appointment was easy to arrange, and later you described several difficulties. Could you help me understand those two parts?” Contradictions may reflect changing contexts rather than error.

Managing emotion and distress

Health research may elicit grief, fear, anger or traumatic memories. The interviewer should acknowledge emotion, offer a pause, remind the participant that they may skip a question and follow the approved referral or safeguarding procedure. Researchers should not promise absolute confidentiality where disclosure may trigger legal or safeguarding duties.

Closing the interview

Ask whether the participant wishes to add anything, summarise key points for correction and explain what happens next. A careful closing reduces the risk that participants leave distressed or uncertain about the use of their contribution.

6. Facilitation of focus groups

At the start, the moderator should establish ground rules: one person speaks at a time, different opinions are welcome, personal stories should not be repeated outside the group and participants may decline to answer. The researcher must explain that confidentiality can be requested but cannot be completely guaranteed in a group setting.

The moderator should balance participation. Dominant speakers may be thanked and then redirected: “I appreciate that view. I would like to hear from people who have had a different experience.” Quiet members can be invited without pressure: “If you are comfortable, how was this for you?” The moderator should avoid turning disagreement into a debate to be won.

Useful techniques include summarising, reflecting contrasting views and asking the group to compare experiences. For example: “We have heard that the appointment system is convenient for some and difficult for others. What might explain that difference?” This encourages interaction rather than a sequence of isolated answers.

Power and status require particular attention. Groups should not combine participants where professional hierarchy, dependency or conflict is likely to suppress speech. If a participant discloses identifiable criticism, the moderator should redirect the discussion away from naming individuals and protect privacy.

7. Sampling and data quality

Qualitative sampling is usually purposive: participants are selected because they can illuminate the phenomenon. Maximum-variation sampling seeks diverse experiences; criterion sampling includes everyone meeting defined conditions; typical-case sampling seeks ordinary cases; and theoretical sampling develops emerging concepts during analysis.

Recruitment should avoid coercion. Recruitment through a treating clinician may make refusal difficult, so the researcher should clarify that participation will not affect care. Compensation should reimburse time and expenses without becoming an undue inducement.

Researchers often use the concept of saturation, but it should be applied carefully. “No new codes appeared” is not identical to “the phenomenon is fully understood.” Researchers should specify whether they mean code saturation, meaning saturation, theoretical saturation or informational redundancy. A saturation rationale should document the sample, analytic process, diversity of participants and evidence that additional data were unlikely to change the interpretation.

Quality can be strengthened through:

• Reflexive field notes.
• Regular team debriefing.
• Comparison of interviews with observations or documents.
• Negative-case analysis.
• Participant clarification where appropriate.
• Transparent documentation of guide changes and sampling decisions.

Member checking is not a universal requirement. Participants may reasonably disagree with a researcher’s interpretation, and findings are analytic constructions rather than transcripts to be “approved.” Its usefulness depends on the research paradigm and the purpose of the check.

8. Recording, transcription and translation

Audio recording generally provides more accurate data than relying on notes alone, but consent must be explicit. A backup recorder may be used when ethically acceptable. Field notes should capture context and initial analytic impressions rather than attempting to reproduce every word.

Transcription decisions affect analysis. Verbatim transcription may include pauses, laughter and overlapping speech when interaction is central. A cleaned transcript may be adequate for content-focused analysis but should not erase meaning relevant to emotion or power. Translation introduces further interpretation. Where possible, retain the original-language transcript, document translator decisions and discuss culturally specific terms with bilingual researchers.

Identifying information should be removed or replaced with pseudonyms. The data-management plan should specify encryption, access controls, retention periods, transfer procedures and secure deletion. Cloud platforms and automated transcription tools should be approved by the institution and assessed for data location, vendor access, confidentiality and accuracy.

9. Remote interviews and online focus groups

Remote methods can include telephone interviews, video interviews and moderated online groups. They may improve access for geographically dispersed participants, people with mobility limitations and staff working shifts. They may also exclude people with poor connectivity, limited digital literacy, hearing impairment, inadequate privacy or restricted data access.

Before the session, test the platform, recording function, headphones, backup communication channel and consent process. Confirm that the participant is in a private location and can end the call safely. For sensitive research, ask whether anyone else is present or able to overhear. Avoid revealing sensitive information in automated calendar invitations or message previews.

Online focus groups require explicit rules about cameras, microphones, chat messages, screenshots and recording. Participants should use private spaces and avoid recording others. The moderator should monitor connection failures and unequal participation. Some participants may contribute through chat, and this should be incorporated into the facilitation plan.

Remote interviews can alter rapport because of reduced non-verbal information. Interviewers should use verbal acknowledgements, explain pauses caused by technology and avoid interpreting a frozen image or delayed response as reluctance. A telephone interview may sometimes be preferable to video when privacy or bandwidth is a concern.

10. Ethics, governance and safety

Ethical review should address informed consent, privacy, confidentiality, risk, voluntary participation, compensation, data security and dissemination. Informed consent is a process, not merely a signed form. Participants should understand the study in accessible language and have opportunities to ask questions.

Confidentiality is especially difficult in small professional communities. Even when names are removed, role, location and unusual events may make a person identifiable. Reports should use carefully selected quotations, aggregate details where necessary and avoid combining descriptors that reveal identity.

The researcher should distinguish confidentiality from anonymity. Anonymity means that identity is not known or cannot reasonably be linked to data; in most recorded interviews, the research team knows the participant’s identity, so confidentiality is the more accurate term.

Researchers also have duties concerning safeguarding, mandatory reporting and immediate risk. These duties must be considered during protocol development, explained during consent and supported by a referral plan. Researchers should not conduct high-risk interviews alone without appropriate safety procedures.

11. Illustrative case: understanding missed outpatient appointments

Background

A regional hospital reported that missed appointments in its diabetes clinic had increased, despite reminder text messages. Management initially proposed sending more reminders.

Problem

The proposed solution assumed that forgetfulness was the main cause. The research team wanted to understand how patients experienced appointment scheduling, transport, communication and care continuity.

Analysis and approach

The team used semi-structured interviews with 24 adult patients purposively sampled across age, residence, employment status and attendance history. It also conducted two focus groups with reception staff and community health workers. Patients were interviewed privately rather than in groups because missed appointments could involve financial hardship or stigma.

The guide began with a narrative question: “Please tell me about the last time you arranged or attended a diabetes appointment.” Probes explored booking, reminders, transport, waiting time, work responsibilities, family support and what happened after a missed visit. The team avoided asking, “Why did you fail to attend?” because that wording implied blame.

Analysis showed several interacting explanations. Some patients did not receive messages because their phone numbers had changed. Others received reminders but could not attend because appointment times changed after shift rosters were issued. A group of patients misunderstood a text abbreviation and believed the appointment was at a different clinic. Community health workers reported that patients who missed one appointment often avoided calling because they expected reprimand. Staff interviews revealed that cancelled clinics were not always communicated consistently.

Outcome

The hospital introduced confirmation messages in plain language, a callback option, flexible rescheduling, a process for updating contact details and a non-judgmental missed-appointment protocol. It also began monitoring cancelled appointments separately from patient non-attendance. The qualitative study did not prove that each intervention would reduce missed appointments, but it identified mechanisms that a reminder-only approach had overlooked.

Learning points

1. Begin with the phenomenon as experienced by participants, not with management’s preferred explanation.
2. Use different methods for different forms of knowledge: patients described lived barriers, while staff explained workflow processes.
3. Avoid deficit language that may produce socially desirable answers.
4. Translate themes into service changes only after considering feasibility and unintended consequences.
5. Combine qualitative findings with operational data when evaluating whether changes improve attendance.

12. Contemporary developments and critical issues

Digital interviewing, encrypted platforms and automated transcription have expanded access, but technology does not eliminate methodological responsibility. Artificial intelligence may assist with transcription, indexing or preliminary coding, yet generated transcripts can misrecognise accents, technical terms, pauses and overlapping speech. AI-generated summaries may also flatten contradiction and remove context. Researchers remain responsible for verification, confidentiality, bias assessment and transparent reporting of tool use.

Participatory and co-designed approaches increasingly involve patients and communities in developing guides, interpreting findings and deciding how results should be used. This can improve relevance and redistribute expertise, but participation must be adequately supported and not treated as unpaid labour.

Intersectionality is also important. Experiences of care may be shaped simultaneously by gender, race, disability, class, migration status, age and professional position. Sampling and analysis should avoid treating categories as isolated variables or assuming that one participant represents an entire group.

Best-practice checklist

Before data collection, confirm that the research question, method, sample, guide, consent process, safeguarding plan and data-management arrangements are aligned. During collection, listen actively, use neutral probes, monitor power and document context. After collection, verify recordings and transcripts, maintain reflexive notes, protect identities and analyse both content and interaction where relevant. In reporting, describe recruitment, setting, interviewer characteristics, guide development, analytic decisions, use of quotations and limitations.

Chapter Summary

Interviews and focus groups are interactive methods for exploring experiences, meanings, practices and organisational processes in health research. Structured interviews maximise consistency, while semi-structured and in-depth interviews provide flexibility and depth. Key-informant interviews illuminate specialised or system-level knowledge but should not be mistaken for population representation. Focus groups generate evidence through participant interaction and are particularly useful for shared norms, service reactions and contrasting perspectives.

A strong guide is derived from the research question, organised into domains, written in accessible open-ended language and supported by purposeful probes. High-quality data collection depends on rapport, listening, silence, neutrality, ethical sensitivity and careful management of power. Sampling should be purposive and justified, while claims about saturation should be specific and modest. Recording, transcription, translation, remote platforms and AI tools introduce additional decisions concerning accuracy, privacy and governance. The central professional responsibility is to produce findings that are credible, ethically obtained and useful without overstating what qualitative evidence can establish.

Key Terms

Interview: A planned interaction in which a researcher elicits and interprets a participant’s account.

Structured interview: An interview using fixed questions and sequence.

Semi-structured interview: An interview guided by topics and flexible questions.

In-depth interview: An interview seeking detailed accounts of experience and meaning.

Key informant: A person with specialised knowledge relevant to the inquiry.

Focus group: A moderated discussion generating data through group interaction.

Moderator: The researcher who guides and facilitates a focus group.

Interview guide: A flexible set of domains, questions and probes.

Probe: A follow-up prompt that clarifies or deepens an account.

Rapport: A respectful working relationship supporting communication and trust.

Reflexivity: Critical examination of how the researcher affects knowledge production.

Purposive sampling: Selecting participants for their relevance to the research question.

Maximum-variation sampling: Seeking participants with diverse characteristics or experiences.

Saturation: A judgment that additional data are unlikely to add relevant conceptual insight.

Confidentiality: Protecting participant information from unauthorised disclosure.

Anonymity: A condition in which identity cannot reasonably be linked to data.

Member checking: Inviting participants to comment on aspects of interpretation or account.

Negative-case analysis: Examining evidence that challenges an emerging interpretation.

Co-production: Developing knowledge through interaction among researchers and participants.

Remote data collection: Conducting interviews or groups through telephone or digital platforms.

Review Questions

Short Answer Questions

1. What is the principal difference between a semi-structured and a structured interview?
2. Why are focus groups not simply several individual interviews conducted together?
3. Define a probe and give one example.
4. Why should confidentiality limits be explained before a focus group begins?
5. What is the difference between confidentiality and anonymity?

Long Answer Questions

1. Explain how a researcher should design and pilot a semi-structured interview guide for a health-services study.
2. Discuss the strengths and limitations of focus groups for researching patient experiences.
3. Explain how rapport, role management and reflexivity influence interview quality.
4. Describe ethical and data-governance issues in remote qualitative interviewing.
5. Critically discuss the use of saturation in qualitative sampling and analysis.

Analytical / Application Questions

1. A hospital wants to understand why patients discontinue hypertension treatment. Design an appropriate combination of interview and focus-group methods, including sampling and key guide domains.
2. During a focus group, a senior clinician dominates discussion while junior staff remain silent. Analyse the methodological problem and propose facilitation responses.
3. A research team uses an automated transcription service for interviews about HIV-related stigma. Develop a governance and quality-assurance plan addressing consent, privacy, transcription accuracy and reporting.

Multiple Choice Questions with Answers

Question 1

Which feature most clearly distinguishes a semi-structured interview from a structured interview?

A. It uses no planned questions.

B. It allows flexible sequencing and follow-up probing.

C. It must be conducted with several participants.

D. It cannot be audio-recorded.

Correct Answer: B. It allows flexible sequencing and follow-up probing.

Question 2

What is the primary source of distinctive data in a focus group?

A. Only the moderator’s observations

B. Only participants’ written responses

C. Interaction among participants as well as their individual accounts

D. A statistically representative sample

Correct Answer: C. Interaction among participants as well as their individual accounts

Question 3

Which question is least leading?

A. How excellent was the new service?

B. Did staff fail to explain the medication?

C. What contributed to the difficulty you experienced?

D. You found the waiting time unacceptable, didn’t you?

Correct Answer: C. What contributed to the difficulty you experienced?

Question 4

Which is an appropriate probe?

A. Was the problem caused by staff shortages?

B. Can you tell me more about what happened next?

C. Why did you make such a poor decision?

D. Wouldn’t an online appointment be better?

Correct Answer: B. Can you tell me more about what happened next?

Question 5

Why might professional hierarchy require separate focus groups?

A. Hierarchy guarantees accurate answers.

B. It can inhibit lower-status participants from speaking freely.

C. It eliminates the need for a moderator.

D. It makes confidentiality automatic.

Correct Answer: B. It can inhibit lower-status participants from speaking freely.

Question 6

Which sampling strategy deliberately seeks participants with diverse characteristics or experiences?

A. Maximum-variation sampling

B. Convenience sampling only

C. Random-digit sampling

D. Census sampling

Correct Answer: A. Maximum-variation sampling

Question 7

Which statement about saturation is most accurate?

A. It always occurs after exactly ten interviews.

B. It means every participant has identical views.

C. It should be justified in relation to the sample, analysis and research purpose.

D. It removes the need for reflexivity.

Correct Answer: C. It should be justified in relation to the sample, analysis and research purpose.

Question 8

What is a major limitation of remote focus groups?

A. They can never generate interaction.

B. They may exclude people with poor connectivity or limited privacy.

C. They cannot use consent procedures.

D. They always produce larger samples than face-to-face groups.

Correct Answer: B. They may exclude people with poor connectivity or limited privacy.

Question 9

Which statement correctly distinguishes anonymity from confidentiality?

A. They are always identical concepts.

B. Confidentiality means identity is unknown to the research team.

C. Anonymity means identity cannot reasonably be linked to the data.

D. Anonymity applies only to focus groups.

Correct Answer: C. Anonymity means identity cannot reasonably be linked to the data.

Question 10

Why should AI-generated qualitative transcripts be checked by researchers?

A. AI systems always delete all recordings.

B. Automated tools may misrecognise accents, technical terms and overlapping speech.

C. Checking transcripts makes consent unnecessary.

D. AI cannot process any spoken language.

Correct Answer: B. Automated tools may misrecognise accents, technical terms and overlapping speech.