Luyện nói tiếng Anh bằng Shadowing qua video: Sequential, Multiple Assignment, Randomized Trial Designs With Dr Kidwell

Hello, and welcome to JAMA Evidence, our podcast series based on core issues in evidence -based medicine.

Today, we will be discussing sequential, multiple assignment, randomized trial designs.

I'm your host, Dr. Roger Lewis, statistical editor for JAMA and co -editor of the JAMA Guide to Statistics and Methods series.

I'm here with Dr. Kelly Kidwell.

Dr. Kidwell, thank you so much for joining us.

Can you go ahead and tell our listeners a little bit about yourself?

Sure.

Thank you so much for having me, Roger.

I'm Kelly Kidwell.

I'm a professor of biostatistics at the University of Michigan School of Public Health, and my expertise is in clinical trial design and analysis,

and in particular, the sequential multiple assignment randomized trials or SMART design.

Well, again, thank you so much for joining us.

So the name of this approach to a trial design is a bit daunting, so let's try to break it up a little bit.

So for each of those terms, can you go ahead and tell us what it means and then how we put it together in designing a clinical trial?

I completely agree with you that the vocabulary around smart design is quite daunting.

So what an excellent first question to start with.

So sequential is about a few things.

One is that we're interested in sequences of treatments.

And in this clinical trial design, we're going to have sequential randomization.

So those terms aren't next to each other, but that's really what's happening and defines this trial,

such that at least some group of participants will be randomized twice in sequence.

So usually there's some upfront randomization

and then a subsequent randomization for some group of participants based on their response to the initial treatment.

And so that multiple assignment comes into the fact that we'll have this sequential treatment assignment.

So individuals could be assigned to multiple treatments.

And this is all wrapped up in a clinical trial.

So we're going to use this SMART clinical trial tool to give us robust evidence for sequences of treatments

that adapt to individual responses or patient status.

So in contrast, in a traditional or a simpler clinical trial design.

We just randomize once, and then the patient either does well

or perhaps doesn't do so well on the treatment to which they are assigned.

And if they don't do well, we often have some sort of rescue therapy or they have to leave the trial.

So how is the experience of a patient or a participant different in a SMART design?

Right.

So in a general, standard, traditional randomized clinical trial, we're really looking at one point in time,

sort of this trial in a vacuum.

The SMART is going to say, let's look at that patient over the course of care.

So in a standard trial where you might stop collecting their data and treating them off trial if they didn't respond, we actually want to use that data.

And in fact, we might want to randomize the treatment to

which rescue therapy they're getting so that we can collect unbiased evidence

and use causal inference to better understand what the effect of that treatment is.

And not just that treatment, but the entire treatment sequence.

And so really for the patient,

it might not be too different than a standard trial because we're going to treat you over the course of your care.

But to the investigators, we're going to make sure that

that data is collected for our trial and there might be randomization for the treatment assignment that follows for that patient.

So, let's think a little bit about how we analyze or what we can learn from those two different approaches.

So in the first, we randomize at a single point in time, and if the patient doesn't do well,

they either drop out of the trial or we assign a rescue therapy.

Whereas in the SMART design, if they require a change in therapy, we actually randomize what we do.

So what's different about what we can learn from those two different approaches?

That's right.

Often in our standard clinical trial, in that one stage where the patient drops out or receives clinician or patient choice therapy,

we no longer make inference or have any decisions or really robust results about that subsequent treatment or their outcome.

Whereas in the SMART, by randomizing that individual and continuing to collect that data,

we will be able to have really robust evidence over the course of care, over the sequence of treatments

that could have adapted based on their response to allow us

to give these sequential guidelines to clinicians to treat the patient not just up front at one point in time,

but over time, and really tailor it to how they're doing, if they responded, if they adhered, if they engaged with that treatment.

So in a sense, this more closely follows what happens in actual clinical care,

where we make a series of treatment decisions based on the patient's prior response, but in this case, you're setting it up

so we can rigorously learn about what is the best next

thing to do based on what's happened to the patient so far.

Is that a fair summary?

Absolutely.

We are looking at a type of clinical trial that more closely mimics the way in which treatment is practiced.

And so that we can provide evidence for these treatment strategies

that adapt to individual's response or patient status in the way that we would actually treat them in practice.

Let's suppose that a SMART design is run and there's two points at which patients are randomized,

maybe the initial therapy and then after some defined period of follow -up.

What does the result look like in the publication that results,

and how can clinicians utilize the results of the SMART in guiding their own clinical care?

One of the amazing advantages of a SMART design is

that it can answer many questions as opposed to just what is the best treatment at one time.

We can think about what is the best treatment to begin with?

What is the best subsequent treatment for those who respond?

What is the best subsequent treatment for those who don't respond?

And then we can think about that entire sequence of, okay, so what do we start with?

How do we change?

When do we change?

For whom do we change and what treatment do we give them?

And so we can really get from SMARTS this sequence of treatments adapted to patient response.

Not only that, we can collect baseline information, so information up front at the start of the trial, and information about the patient and how they're doing with the treatment,

how they're adhering, how they're engaging, if they're having side effects,

and use that information to better tailor the treatment assignment to that individual so that after the trial, right, we have all this information.

And we can look at what are the best sequences of treatments for which particular individuals.

The fact that the SMART design more closely matches actual clinical care,

which the sequences of therapies we offer patients is informed by their responses to prior therapies is an obvious advantage.

But what are some of the limitations or challenges in either the implementation or the interpretation of a SMART clinical trial?

I think there are a few limitations and also just lack of knowledge of this design and its opportunities that it offers.

Some of the limitations are that individuals think, oh, now that we have multiple randomizations or multiple assignments in sequence,

this is now logistically infeasible, or you're going to lose more people if you have to consent multiple times.

So one thing is that we often just consent once at the beginning of the smart design.

and we may randomize multiple times.

Alternatively, you could randomize up front to that sequence that they may follow, whether they respond or do not respond.

So logistically, this doesn't have to be very difficult.

We also have lots of technology that helps us with complex trials, if you want to say that it's complex, although it's not.

Number two is that individuals are worried about bias coming into play from a smart design

because of these multiple randomizations and the subsequent randomization depending upon patient response or status.

100

This is not a concern because we do not compare these treatment pathways or you first get A,

101

you respond and you continue versus you first get A, you do not respond and you switch to B.

102

We're not going to compare those people.

103

We know that they're different.

104

One responded and one didn't, and we can have some issues if we do those comparisons.

105

We're looking at what we call adaptive interventions or dynamic treatment regimens that include that entire treatment sequence that's adapted.

106

So start with A, and if you respond, continue.

107

And if you don't, receive B.

108

We're going to think about what are the effects of that entire sequence that's tailored to the response.

109

Another misconception is that a SMART is an adaptive design.

110

Usually a SMART is a fixed design, not an adaptive design, meaning

111

that we're not going to use information from patients who have

112

been through the trial to change the trial in some way for future patients.

113

So we're not going to change randomization probabilities or eligibility criteria based on the information we've gained from the trial thus far.

114

What does happen in a SMART is that patients' treatment might change based on their own response.

115

So there might be a within -person adaptation, but there's no between -person adaptation.

116

And so generally, these aren't adaptive designs.

117

Since we're trying to learn more from a clinical trial that's a smart design versus a traditional design

118

that just has a single point of randomization, it seems like we might need larger sample sizes to answer the additional questions.

119

What are the challenges, if any, with respect to sample size requirements for a smart design?

120

If you see pictures of smart designs, they look a little bit intimidating because there's a fair amount of branching that occurs due to these sequential randomizations.

121

And often investigators are quite worried that we're going to require thousands of patients.

122

The thing is that to size a SMART is just like sizing any other trial.

123

And so it's really dependent on the primary aim, which can be similar to a standard one -stage trial and compare,

124

for example, the initial treatments.

125

Alternatively, the aim can be based on those embedded tailored sequences of treatments or the adaptive interventions.

126

And for those perhaps more patients may be required than a standard trial.

127

However, this really isn't an apples -to -apples comparison of a SMART design to a standard or traditional one -stage trial.

128

As I mentioned before, the questions

129

that we can answer in a SMART are much different and many more than we could with a standard design.

130

And so perhaps we may need more depending on the primary aim, but we're going to get a lot more information,

131

data, robust results, and more practical information that we can use after a smart design.

132

In your article, you discuss a study that was performed by Fortney et al that used a smart approach.

133

Can you tell us a little bit about what they did, what they found, and how they analyzed their data?

134

Sure.

135

So in this study, they had a design where individuals were sequentially randomized to better understand some effects of direct

136

and indirect telecare.

137

And what they did is they answered questions that are sort of more similar to one stage design.

138

So they said, okay, so what's the best first thing to start with?

139

And then subsequently, what do we do if individuals aren't responding well to that initial treatment or intervention.

140

So they haven't yet pieced together that full sequence.

141

It seems like that's subsequent analyses for this group in terms of figuring out,

142

so what is the best overall sequences of treatment that can adapt to the individuals?

143

While they answered those individually, what they could find when they move forward with the analysis of the adaptive intervention is

144

that there might be interactions between the initial treatment

145

and the subsequent treatment for

146

which we can only really quantify in these types of SMART designs as opposed to individual randomized clinical trials.

147

And so what they likely will subsequently do is look at

148

that sequence of treatment over the course of the trial

149

or those adaptive interventions to see if there are any interactions between the first step and the second step.

150

I'm Roger Lewis, and I've been speaking with Professor Kelly Kidwell about sequential multiple assignment randomized trials.

151

You can find a link to her article in this episode's description.

152

This episode was produced by Daniel Musisi at the JAMA Network.

153

To follow this and other JAMA Network podcasts, please visit us online at jamanetworkaudio .com.

154

Thanks for listening.

📺 Cùng kênh

Balloon Angioplasty for Intracranial Atherosclerotic Stenosis

IELTS PrepTOEFL PrepUniversity Students

✨ Video gợi ý

28 câu

Explanation of Vaccine Adjuvants

IELTS PrepTOEFL PrepUniversity Students

22 câu

Journal Highlight: 30-Day Outcomes of Resolute Onyx Stent for Symptomatic Intracranial Stenosis

IELTS PrepTOEFL PrepUniversity Students

623 câu

What Happens When Science Clashes with the Public?: Crash Course Scientific Thinking #7

IELTS PrepTOEFL PrepUniversity Students

36 câu

Neurovascular Specialists Discuss Stroke Treatments

IELTS PrepTOEFL PrepUniversity Students

75 câu

What AI can (and can't) currently do in a real lab | Jennifer Doudna

IELTS PrepBusiness EnglishUniversity Students

64 câu

How To Get Away with Murder - "Day One and you're unprepared"

IELTS PrepTOEFL PrepCambridge Exam Prep

Từ vựng và ghi chú luyện nói cho bài học này

Bài luyện nói này dựa trên video “Sequential, Multiple Assignment, Randomized Trial Designs With Dr Kidwell”. Những từ được nhắc lại nhiều nhất trong bài: treatment, trial, design, smart, patient. Video này có 103 câu và 2153 từ để luyện shadowing. Phần lời nói dài 13:26. Người nói nói với tốc độ tự nhiên, khoảng 160 từ mỗi phút, gần với hội thoại hằng ngày. 84% số từ nằm trong 3.000 từ tiếng Anh thông dụng nhất; phần còn lại nên xem trước khi luyện. 18 trong 103 câu là câu hỏi, nên bài này hợp để luyện ngữ điệu câu hỏi và cách trả lời ngắn.

Từ vựng quan trọng trong video

15 từ đáng học trong video, kèm phiên âm và nghĩa:

Từ	Phiên âm	Nghĩa
sequence danh từ	/ˈsiː.kwəns/	dãy
collect động từ	/ˈkɑlɪkt/	thu thập, gom góp
tailor danh từ	/ˈteɪlɚ/	thợ may
rescue động từ	/ˈɹɛs.kjuː/	cứu
compare động từ	/kəmˈpɛɚ/	so sánh, so
participant danh từ	/pɑːɹˈtɪsəpənt/	người tham gia
consent động từ	/kənˈsɛnt/	đồng ý
versus	/ˈvɝsəs/	đối, đối đầu
depend động từ	/dɪˈpɛnd/	phụ thuộc
oppose động từ	/əˈpoʊz/	chống đối
analyze động từ	/ˈæn.əˌlaɪz/	phân tích
investigator danh từ	/ɪnˈvɛs.tɪˌɡeɪ̯.təː/	người điều tra
occur động từ	/əˈkɜː/	xảy ra
summary danh từ	/ˈsʌm.ə.ɹi/	tóm tắt
requirement danh từ	/ɹɪˈkwaɪɹmənt/	yêu cầu

Những câu đáng luyện lại

Các câu ngắn, trọn ý trong video mà bạn có thể dùng lại khi giao tiếp hằng ngày:

Let's try to break it up a little bit.
One is that we're interested in sequences of treatments.
We can think about what is the best treatment to begin with?
How do we change?
When do we change?

Ngữ pháp trong video

Những cấu trúc người nói dùng nhiều nhất, kèm đúng cụm từ trong video:

Cấu trúc	Trong video
Câu bị động be + quá khứ phân từ — nhấn vào việc xảy ra, không phải người làm	will be randomized · could be assigned · are assigned
Mệnh đề quan hệ who / which + mệnh đề — thêm thông tin về người hoặc vật	those who respond · those who don · aim, which can

Phát âm cần chú ý

Người nói dùng 44 dạng rút gọn, ví dụ we're, they're, I'm. Hãy nói theo dạng ngắn đúng như bạn nghe.

Âm “sh” và “zh”: intervention /ɪntɚˈvɛnʃən/, limitation /lɪmɪˈteɪʃən/, interaction /ˌɪn.təˈɹæk.ʃən/, adaptation /ˌæd.æpˈteɪ.ʃən/, contrast /kənˈt(ʃ)ɹæst/
Từ dài — đặt trọng âm cho đúng: intervention /ɪntɚˈvɛnʃən/, participant /pɑːɹˈtɪsəpənt/, limitation /lɪmɪˈteɪʃən/, interaction /ˌɪn.təˈɹæk.ʃən/, subsequently /ˈsʌb.sɪ.kwənt.li/

Những âm người Việt thường gặp khó:

Cụm phụ âm cuối — đọc đủ từng âm, đừng nuốt âm cuối: sequence /ˈsiː.kwəns/, assignment /əˈsaɪn.mənt/, subsequent /ˈsʌbsɪkwənt/, adapt /əˈdæpt/, collect /ˈkɑlɪkt/
Âm /tʃ/, /dʒ/ và /ʒ/ — tiếng Việt không có, đừng đọc thành “ch” hay “gi”: engage /ɪnˈɡeɪd͡ʒ/, individually /ˌɪndɪˈvɪd͡ʒ(u)əli/, eligibility /ˌɛl.ɪ.d͡ʒəˈbɪl.ɪ.ti/, randomization /ˌɹændəməˈzeɪt͡ʃən/, sequential /səˈkwɛntʃəl/
Âm /æ/ — mở miệng rộng hơn âm “e”: adaptive /əˈdæp.tɪv/, adapt /əˈdæpt/, whereas /(h)wɛɹˈæz/, interaction /ˌɪn.təˈɹæk.ʃən/, podcast /ˈpɑdˌkæst/

Cách luyện với video này

Nghe hết video một lần, chưa cần nói, và ghi lại những từ bạn chưa biết.
Bắt đầu ở tốc độ 0,75×, nói đuổi từng câu, rồi quay lại tốc độ bình thường khi đã quen.
Ghi âm giọng mình rồi so với bản gốc, chú ý các từ như sequence, collect, tailor.

Phương Pháp Shadowing Là Gì?

Shadowing là kỹ thuật học ngôn ngữ có cơ sở khoa học, ban đầu được phát triển cho chương trình đào tạo phiên dịch viên chuyên nghiệp và được phổ biến rộng rãi bởi nhà đa ngôn ngữ học Dr. Alexander Arguelles. Nguyên lý cốt lõi đơn giản nhưng cực kỳ hiệu quả: bạn nghe tiếng Anh của người bản xứ và lặp lại to ngay lập tức — như một "cái bóng" (shadow) đuổi theo người nói với độ trễ chỉ 1–2 giây. Khác với luyện ngữ pháp hay học từ vựng bị động, Shadowing buộc não bộ và cơ miệng phải đồng thời xử lý và tái tạo ngôn ngữ thực tế. Các nghiên cứu khoa học xác nhận phương pháp này cải thiện đáng kể phát âm, ngữ điệu, nhịp điệu, nối âm, kỹ năng nghe và độ lưu loát khi nói — đặc biệt hiệu quả cho người luyện IELTS Speaking và muốn giao tiếp tiếng Anh tự nhiên như người bản ngữ.

Phương pháp shadowing: đọc hướng dẫn từng bước đầy đủ →

Luyện nói tiếng Anh bằng Shadowing qua video: Sequential, Multiple Assignment, Randomized Trial Designs With Dr Kidwell

Balloon Angioplasty for Intracranial Atherosclerotic Stenosis

Explanation of Vaccine Adjuvants

Journal Highlight: 30-Day Outcomes of Resolute Onyx Stent for Symptomatic Intracranial Stenosis

How evolution works, in 54 minutes | Sean B. Carroll: Full Interview

Personalised Cancer Care - Prof Mark Middleton

Other Corneal Disorders - CRASH! Medical Review Series

Evidence Markets: Upgrading Prediction Markets

Evidence and challenges in MeVO and distal occlusions, what do we know in 2026?

Anna Lembke on Pleasure, Pain, and Recovery in a Dopamine-Driven World

What Happens When Science Clashes with the Public?: Crash Course Scientific Thinking #7

Neurovascular Specialists Discuss Stroke Treatments

What AI can (and can't) currently do in a real lab | Jennifer Doudna

How To Get Away with Murder - "Day One and you're unprepared"

Từ vựng và ghi chú luyện nói cho bài học này

Từ vựng quan trọng trong video

Những câu đáng luyện lại

Ngữ pháp trong video

Phát âm cần chú ý

Cách luyện với video này

Phương Pháp Shadowing Là Gì?