(CNN) -- Long-awaited data from an international study have shown no evidence of increased risk of brain tumors associated with mobile phones, except in people who have the most exposure.
But design flaws of the Interphone study, which is partly industry funded, suggest that the latest results cannot be taken to mean that cell phones and brain cancer are unrelated, critics say.
"I'm not telling people to stop using the phone. I'm saying that I can't tell you if cell phones are dangerous, but I can tell you that I'm not sure that they are safe," said Dr. Devra Davis, professor of preventive medicine at Mount Sinai Medical Center in New York.
The study itself, to be published Tuesday in the International Journal of Epidemiology, acknowledged that the findings were not definitive and called for more research.
But Daniel Krewski, professor of epidemiology at the University of Ottawa in Ontario and one of the Interphone collaborators, said the study went to "great lengths to make sure that the results were scientifically sound."
The results showing no connection between cancer and phones are "reassuring," he said. "It tells us that we don't have an epidemic of brain cancer on our hands associated with mobile phones."
At the highest exposure levels -- using a mobile phone half an hour a day over a 10-year period -- the study found a 40 percent increased risk of glioma brain tumors. With adjustments for statistical biases, that turned into 80 percent. But Krewski and colleagues say that there is not enough evidence to show a causal connection, and the group of participants using their phones this much was relatively small.
In regular mobile phone users, the study found that phone radiation exposure actually decreased the likelihood of tumor, although this "protective" effect may have resulted from limitations in the study, the authors wrote.
But critics say that "high" exposure may actually be most representative of the average cell phone user, meaning the study may underestimate the actual risk of tumors.
Five ways to reduce cell-phone radiation
The decade-long study defines a "regular" cell phone user as someone who uses a phone at least once a week for at least months. Results were based on patients' average talk time ranging from 120 minutes to 150 minutes a month -- some U.S. users talk that much on the phone in a single week, Davis said.
Another concerning flaw in the study's methodology was that it did not address cordless phone usage, which has been found in other studies to have a risk of tumors similar to cell phones, said Lloyd Morgan, a retired engineer and consumer advocate who has published several analyses on the phone-cancer issue. The study therefore may underestimate the risk of brain tumors because it does not take this exposure into account.
Children and young adults were not included in the study, critics note, despite other findings that exposure beginning at a younger age is associated with increased risk of brain tumors, Morgan said. Interphone also released results on only two types of brain tumors: meningioma and glioma.
Previous Interphone research on cell phone users' ability to recall how many hours they spend on the phone has found that heavy users generally underestimate that time, and light users overestimate it, Morgan said. This suggests that the current study, which used people's recollections of phone usage, underestimates the risk of tumors in many users, he said.
But Krewski said that overall people generally have a good ability to recall the number of phone calls reasonably accurately, and tend to overestimate call duration by about 40 percent.
The way that the study selected a control group also biases the results, Davis said.
"Their comparison isn't between people who used cell phones and didn't. It was between people who used their phone less than once a week, and more than once a week, all on self-reflection memory of their mobile use, the day after they've had a brain operation," she said.
Krewski said the authors acknowledge this limitation of participant selection, and corrected for it when doing statistical analyses.
The study was conducted at 16 locations in 13 countries, not including the United States. In hospital interviews, brain tumor patients were asked to answer a series of questions based on memory about their mobile phone use habits prior to being diagnosed.
They were asked to reflect on how many hours they talked on their phone a month, how long they had been using their cell phone, if they remembered how many phone calls they made a month and to what side of the head they typically held the phone.
Data collection took place between 2000 and 2004, and results have been expected since 2005. Tuesday's study is still not a reflection of the entire data set.
Cell phone popularity has gone up since the data were gathered, Davis noted.
The World Health Organization's International Agency for Research on Cancer is coordinating Interphone. Several of the researchers analyzing the data acknowledged that they received money from the mobile phone industry.
An analysis in the Journal of Clinical Oncology last year found that studies that have looked at people who had used cell phones 10 years or longer tended to find the strongest risk of tumors. Eight of the strongest studies came from the same researcher: oncologist Dr. Lennart Hardell in Sweden. This research made sure the investigators did not know which participants had tumors when they conducted the interviews about cell phone use, and they did not receive funding from industry groups.
But there may be something unique to Sweden in terms of cancer susceptibility and cell phone use, experts say, and therefore Hardell's findings may not be widely applicable.
The Interphone study has not been released in its entirety. There are ongoing analyses of cell phones' effects on other kinds of tumors, and other analyses. There is also a study to correlate the position of where participants' tumors started to grow in their brains, and what the energy level from cell phones are at that area in the brain.
CNN's Danielle Dellorto contributed to this report.