Though lots of people used the expensive government health insurance portal to get information on the Affordable Care Act, also called Obamacare, far fewer could successfully use it to sign up.

As the stories of its flaws mounted, larger percentages instead talked to call centers or a navigator without using the website at all. That's a win for the government, which needed to show some success after expending a great deal of political capital and taxpayer money, but not without cost. How much cost is unclear, the government has not yet disclosed how many people signed up and actually paid anything, or how many stopped when they discovered that on top of payments they now had $4,000 a year deductibles.

With a new enrollment period just weeks away, Democrats are scrambling to understand how to keep the initial promise alive. A new report by the Episcopal Health Foundation and Rice University's Baker Institute for Public Policy may give them some insight, at least into how people will sign up. Even after fixes, only 70 percent of respondents believe the website is good, which means the costs are going to stay high.

Marks said the Texas survey results that found personalized service most helpful are supported by national results showing people assisted by enrollment professionals were more likely to enroll in coverage - that may be some of the confusion. People have no choice but to enroll in coverage, and if the government told more people that, there would be higher uptake. Unfortunately, there would also be higher votes for opponents of Democrats next month if they made that more clear.

No matter which enrollment method they tried, many Texans found it difficult to determine whether they were eligible for a subsidy under the ACA, the report showed. Without that information, consumers can't make informed decisions on whether to purchase a plan. The difference in the price of a subsidized plan versus a non-subsidized plan can be hundreds of dollars each month and that is without considering the high deductibles.

"This is an important step because the cost of a plan depends on the amount of subsidy available," said Vivian Ho, the chair in health economics at Rice's Baker Institute, a professor of economics at Rice and a professor of medicine at Baylor College of Medicine. "We know from previous research that many who were eligible for a subsidy didn't purchase a plan. If clearer eligibility and financial assistance information had been available, more people might have enrolled in coverage."

Or they found that the plan was too expensive. A subsidized plan with a deductible so high people can't afford to use the plan won't get purchased, they would rather pay the penalty.

The majority of Texans who used the website said the top way to improve the process would be to have better information available to determine eligibility for financial assistance. For those who used the call center to try and get cheaper insurance, they complained about wait times.

"It's important to understand what Texans found most effective and where improvements are needed," said Elena Marks, CEO of the Episcopal Health Foundation and a nonresident fellow in health policy at the Baker Institute. "With the second enrollment period just weeks away, it's important for each enrollment method to be at peak performance to help the hundreds of thousands of Texans who are eligible for subsidized health insurance plans, but remain uninsured."

The report is the ninth in a series on the implementation of the ACA in Texas co-authored by Marks and Ho.

The Health Reform Monitoring Survey (HRMS)-Texas report is based on a national project that provides timely information on implementation issues under the ACA and changes in health insurance coverage and related health outcomes. The Episcopal Health Foundation and Baker Institute are partnering to fund and report on key factors about Texans obtained from an expanded representative sample of Texas residents. Today's report contains responses from 1,595 Texans in September 2013 and 1,538 in March 2014.

The survey was developed by the Urban Institute, conducted by the company GfK and jointly funded by the Robert Wood Johnson Foundation, the Ford Foundation and the Urban Institute.

The analyses and conclusions based on HRMS-Texas are those of the authors and do not represent the views of the Urban Institute, the Robert Wood Johnson Foundation or the Ford Foundation.