The largest study ever done on treating depression has found that patients who didn't get well with the first medicine they tried had a good chance of succeeding the second time around.

Up to one-third of those who added or changed medicines recovered from the crushing illness that is America's top mental health problem, researchers said.

This is good news by itself, but the bigger picture is even more encouraging, doctors say. When viewed with earlier results, the new findings mean that roughly half of people who suffer from depression can get over it — not just improve their symptoms — with adequate medication.

"The goal here was to find treatments that help people to get well, not just better," said Dr. Thomas Insel, director of the National Institute of Mental Health. "We have safe and effective treatments."

His agency paid for the $35 million study, which involved thousands of people across the United States and has been widely praised as a real-world test of popular drugs that have received only limited testing until now.

The study found little difference among the five drugs tested — Celexa, Zoloft, Wellbutrin, Effexor and Buspar — and wasn't designed to compare them. All proved similarly effective and relatively safe. The clear message, doctors said, was that antidepressants should be given a 6-to-12-week chance to work and that if one doesn't help, another should be tried.

"It's important not to give up if the first treatment doesn't work fully," or causes side effects, said one study leader, Dr. John Rush of the University of Texas Southwestern Medical Center in Dallas.

Two reports from the study were published Thursday in the New England Journal of Medicine.

About 15 million Americans each year suffer depression, which so often recurs that doctors sometimes talk of it as an emotional cancer that is put "in remission" rather than cured.

"We're talking about a very real public health challenge," Insel said. "This is the leading cause of disability in Americans ages 15 to 44," not just a case of "the blues," he said.

Nearly two dozen antidepressants are on the market; 189 million prescriptions were filled last year alone. Evidence on their effectiveness is limited, and the government recently ordered stronger warnings that some can worsen suicidal tendencies in teenagers in rare cases. The risk in adults is still being studied.

The big federal study first tested Forest Laboratories' Celexa, a newer type of antidepressant called a selective serotonin reuptake inhibitor, or SSRI, mostly because it's an easy-to-take daily pill.

One-third of the roughly 3,000 taking it recovered, though they generally took higher doses and were monitored more closely than most patients, researchers reported several months ago.

The new research, step 2 of the study, was on people who didn't get well the first time around. They had depression for 16 years on average and two-thirds had other mental or physical problems.

Out of this group, 727 chose to switch from Celexa to a different medication and were randomly assigned to get either Zoloft, another SSRI made by Pfizer Inc.; Wellbutrin, a non-SSRI antidepressant made by GlaxoSmithKline; or Effexor, an antidepressant made by Wyeth that works on another brain chemical in addition to the one targeted by SSRIs.

Roughly one-fourth became symptom-free within 14 weeks.

Another 565 patients chose to add a second drug to Celexa and were given either Wellbutrin or Buspar, a Bristol-Myers Squibb anti-anxiety medication that can boost the effectiveness of SSRIs.

Within 14 weeks, about one-third were symptom-free. Those on Wellbutrin had slightly fewer symptoms and side effects than those on Buspar.

The study will continue to test third and even fourth treatment attempts, and to analyze genes to see if any patterns emerge with particular drugs.

"It's quite possible in the near future we may be able to predict who's going to respond to what," said another study leader, Dr. Madhukar Trivedi of UT Southwestern.

In an editorial in the New England journal, Dr. David Rubinow of the University of North Carolina at Chapel Hill wrote that the study is encouraging, because half got well on drugs, but discouraging, because half did not.

Roughly four out of 10 people in the study were unemployed and nearly that many had no health insurance. Without access to treatment and less societal stigma toward depression, millions will continue to suffer, he wrote.