Drug Effectiveness: The Missing Number Behind Every Prescription

You might assume that drug effectiveness would be one of the most important things a doctor would discuss when writing a new prescription. That almost never happens. Patients may be told what a medicine is supposed to do and what serious side effects to watch out for. They are rarely told the actual odds that the drug will help them.

That missing information may be one of the biggest blind spots in modern medicine. Without understanding drug effectiveness in clear numerical terms, patients cannot make truly informed decisions about treatment. They may be agreeing to take a medicine without knowing whether the likely benefit is substantial, modest or barely better than a placebo.

Consumer Shopping Is Scientific

How do you buy stuff? In particular, how do you make decisions about big purchases such as for a car, TV, microwave oven or cell phone? Many people check for ratings by other buyers or rely on evaluations by nonprofit organizations such as Consumer Reports (CR). Effectiveness, cost, reliability and safety are key measurements.

If you have used CR for your own shopping decisions, you know that the information is presented in easy-to-understand tables. There is a lot of data crammed into them. In one recent review of air filters, there were 18 columns of data. You get understandable ratings and comparative scores.

Of course, CR is especially renowned for its car reviews. Each vehicle is rated with a number so you can easily compare one model to another. They also analyze a wide range of criteria, from acceleration and braking to predicted reliability and overall miles per gallon or EV range. And let’s not forget a key metric: “OWNER SATISFACTION“! Wouldn’t you like to be able to assess your prescription medicines with a somewhat similar data grid?

Drug Effectiveness Is a Well-Guarded Secret

Most physicians and patients have no idea how effective one medicine might be compared to another. Rarely are people told how likely they are to experience side effects. Patients are almost never informed in any quantitative way how well a medicine they are prescribed will actually reduce their risk of a serious health problem.

People must rely on the prescriber, who may have little understanding about the actual effectiveness of the medicine being recommended. Search for drug information on the most popular websites and you will be told a little something about how it works, which medications might interact, common side effects, serious side effects and how to take it. Rarely, if ever, will you be given meaningful information about drug effectiveness.

When You Get a New Prescription

What happens when your doctor recommends a blood pressure medicine, an antidepressant or a cholesterol-lowering drug?

Where is the scorecard showing:

• How likely the medicine is to help?
• How likely it is to cause harm?
• How much better it works than placebo?
• How it compares with alternatives?

For most prescription medicines, that information is hard to find.

The Most Important Number You Have Never Heard Of: NNT

Has your physician, nurse practitioner (NP), or physician’s associate (PA) ever discussed the Number Needed to Treat or NNT? What about your pharmacist? I didn’t think so.

The NNT tells you how many people need to take a medicine for one person to benefit. The lower the number, the better!

An NNT of 1 would mean everyone benefits! That almost never happens. Yet most patients assume their medicine works far better than it actually does based on randomized controlled trials.

Drug Effectiveness for Lipitor

As you no doubt know if you read this newsletter, statins are the most prescribed drugs in this country. At last count, over 50 million people in the US take a statin to reduce their risk of a heart attack or stroke. Lipitor (atorvastatin) alone is taken by about 30 million Americans.

Many years ago the company that sold Lipitor advertised it this way:

“In patients with multiple risk factors for heart disease,

“Lipitor reduces risk of heart attack by 36%*

“If you have risk factors such as family history, high blood pressure, age, low HDL (‘good’ cholesterol) or smoking

“*That means in a large clinical study 3% of patients taking a sugar pill or placebo had a heart attack compared to 2% of patients of patients taking Lipitor.”

If you did not read the smaller print after the asterisk you may have missed the 1% absolute risk reduction between placebo and Lipitor.

Statin Stats (

If you go to the website titled TheNNT.com and put statins into the search engine at the top of the page you will discover 4 options:”

I am guessing here, but I suspect that the vast majority of patients on statins are those with no prior heart disease. They may have elevated LDL-cholesterol, but no prior heart attacks, strokes or diagnosed atherosclerosis.

According to the analysis by TheNNT.com researchers:

People who take statin-type cholesterol-lowering medicines for 5 years would have an NNT of 104. In other words, 104 people would need to take a statin for 1 to avoid a heart attack. 154 people would need to take a statin for 1 to avoid a stroke.

People who have “known heart disease” get more benefit. TheNNT.com analysis reports that people who have had a heart attack or received a stent get the following benefits after 5 years of statin treatment:

• “1 in 83 were helped (life saved)
• “1 in 39 were helped (preventing non-fatal heart attack)
• “1 in 125 were helped (preventing stroke)”

If that seems confusing to you, TheNNT.com offers percentages:

For people with “known heart disease” statins provided the following numbers:

• “96% saw no benefit
• “1.2% were helped by being saved from death
• “0.8% were helped by preventing a stroke”

We wonder how many physicians realize that after 5 years of statin treatment, the majority of patients would not see much tangible benefit. That is why absolute risk reduction is a much better reflection of what would happen in real life compared to relative risk reduction.

Ask your doctor for the NNT the next time she prescribes any medicine. She can calculate that number by putting 100/absolute risk reduction.

Many clinical trials provide the absolute risk reduction but the prescriber will need to search for it to calculate the NNT. You may want to ask your favorite artificial intelligence search engine what the NNT is for DRUG X. You may be surprised to learn how many people have to take a particular drug for a long time for one person to benefit.

Drug Effectiveness? When Negative Studies Disappear

Doctors depend on published medical studies to guide prescribing. But what if some of the studies never see the light of day? That happens more often than many people realize.

Consider the benzodiazepine alprazolam (Xanax). This anti-anxiety medicine has long been considered highly effective for panic disorder.

But an analysis in Psychological Medicine, Oct. 19, 2023 uncovered something disturbing.

Researchers found five FDA-registered clinical trials for extended-release alprazolam. Only three were published. And according to FDA reviewers, only one of the five clearly showed a positive outcome.  That means the published literature created a much stronger impression of effectiveness than the full FDA record.

This is called publication bias.

Positive studies get published. Negative studies often disappear. That distorts the evidence.

The authors of this study conclude:

“We found that alprazolam XR may be less effective than the published literature would suggest. According to the published literature, every trial of alprazolam XR found it to be effective. By contrast, according to the FDA, only one of five trials was positive.

“This study brings to light unpublished trial data and provides a more balanced and realistic view of the efficacy of alprazolam XR, compared to what has been previously reported. It is unknown whether the discrepancy between FDA and journal trial data is greater or smaller for other benzodiazepines. This adds to the literature on publication bias in clinical trials for drugs for psychiatric conditions, including major depressive disorder, bipolar disorder, anxiety disorders, and schizophrenia.”

If the evidence is distorted, so are treatment decisions.

Drug Effectiveness Data and Antidepressants

This problem became painfully clear with antidepressants.

In 2008, psychiatrist Erick Turner and his colleagues reviewed 74 antidepressant studies submitted to the U.S. Food and Drug Administration (New England Journal of Medicine, Jan. 17, 2008).

Nearly one-third had never been published. When studies were positive, they were published 97 percent of the time. When studies were negative, only 8 percent were published as negative. That could create a deeply misleading scientific record.

Studies have shown that placebo response rates average 31-45%, compared to 50% with antidepressants (Walsh et al, JAMA, April 10, 2002; Stolk et al, Annals of Pharmacotherapy, Dec. 2003). One meta-analysis concluded that the placebo response rate was as high as 82% (Kirsch et al, PLoS Medicine, Feb. 2008).

When researchers reanalyzed the complete FDA data, including unpublished studies, they found that several antidepressants were barely more effective than placebo for mild to moderate depression.

A more recent review by one of the most prestigious organizations (Cochrane Database of Systematic Reviews, May 24, 2021) reviewed the effectiveness of antidepressants for “major depressive disorders in children and adolescents.”

The conclusions about antidepressant drug effectiveness:

“Most antidepressants may be associated with a “small and unimportant” reduction in depression symptoms on the CDRS‐R scale (range 17 to 113) compared with placebo…Findings suggest that most newer antidepressants may reduce depression symptoms in a small and unimportant way compared with placebo.”

That does not mean antidepressants never work. They clearly help some people. But patients deserve realistic expectations. That’s because these drugs can cause nausea, insomnia, sexual dysfunction, dizziness and weight gain, among many other side effects.

The question should never be simply:

“Does this drug work?”

It should be:

• “How often does it work?”
• “How well does it work?”
• “For whom is it likely to work well?”
• “What is the NNT?”
• “How likely is it to cause harm?” (NNH or number needed to harm)

Put simply:

• “How many people need to take this medicine before one person benefits?”

Final Words:

Americans can compare cars, refrigerators and coffee makers with extraordinary precision.

But when it comes to medicines, products that can change lives or end them, they often get less useful information than when buying a toaster. Until both prescribers and patients adopt an objective measure of drug effectiveness, people won’t be able to make informed decisions about their health.

Before filling your next prescription, ask your doctor some simple questions.

What are the actual odds this medicine will help me?

You could follow that with this request: Please show me the NNT for Drug X. 

Put another way, you could ask: if 100 people take this medicine, how many would be expected to get a detectable benefit?

People frequently ask their surgeons how likely a particular procedure is to be helpful. There is no reason a prescriber should not be asked how likely a medicine will produce a beneficial result. Your doctor, nurse practitioner, physician’s associate or pharmacist may not have an answer at their fingertips, but they should be able to research it. Artificial intelligence can actually calculate this sort of thing in seconds if you ask about the NNT…such as: “What is the NNT for Jardiance?”

Was this article of value? If so, please vote in the upper left hand corner. While there, you could click on one or more of the social media icons and send it to a friend or family member via email. Most people have never considered drug effectiveness very important. Now you know that it is crucial!

If you have a medication experience you would like to share, please scroll to the bottom and include it in the comment section. Thank you for supporting our work.