Twelve years ago, the U.S. government introduced a powerful new tool to help people make a wrenching decision: which nursing home to choose for loved ones at their most vulnerable. Using a simple star rating — one being the worst, five the best — the system promised to distill reams of information and transform an emotional process into one based on objective, government-blessed metrics.
The star system quickly became ubiquitous, a popular way for consumers to educate themselves and for nursing homes to attract new customers. During the coronavirus pandemic, with many locked-down homes unavailable for prospective residents or their families to see firsthand, the ratings seemed indispensable.
But a New York Times investigation, based on the most comprehensive analysis of the data that powers the ratings program, found that it is broken.
Despite years of warnings, the system provided a badly distorted picture of the quality of care at the nation’s nursing homes. Many relied on sleight-of-hand maneuvers to improve their ratings and hide shortcomings that contributed to the damage when the pandemic struck.
More than 130,000 nursing-home residents have died of COVID-19, and The Times’ analysis found that people at five-star facilities were roughly as likely to die of the disease as those at one-star homes.
The ratings program, run by the U.S. Centers for Medicare & Medicaid Services, relies on a mix of self-reported data from more than 15,000 nursing homes and on-site examinations by state health inspectors. Nursing homes receive scores based on how they fare in those inspections; how much time nurses spend with residents; and the quality of care that residents receive. Those three grades are then combined into an overarching star rating for each nursing home.
To evaluate the ratings’ reliability, The Times built a database to analyze millions of payroll records to determine how much hands-on care nursing homes provide residents, combed through 373,000 reports by state inspectors and examined financial statements submitted to the government by more than 10,000 nursing homes.
The Times obtained access to portions of the ratings data that aren’t publicly available from academics who had research agreements with the Centers for Medicare & Medicaid Services, or CMS.
Among The Times’ findings:
— Much of the information submitted to CMS is wrong. Almost always, that incorrect information makes the homes seem cleaner and safer than they are.
— Some nursing homes inflate their staffing levels by, for example, including employees who are on vacation. The number of patients on dangerous antipsychotic medications is frequently understated. Residents’ accidents and health problems often go unreported.
— In one sign of the problems with the self-reported data, nursing homes that earn five stars for their quality of care are nearly as likely to flunk in-person inspections as to ace them. But the government rarely audits the nursing homes’ data.
— Data suggest that at least some nursing homes know in advance about what are supposed to be surprise inspections. Health inspectors still routinely found problems with abuse and neglect at five-star facilities, yet they rarely deemed the infractions serious enough to merit lower ratings.
At homes whose five stars masked serious problems, residents developed bed sores so severe that their bones were exposed. Others lost the ability to move.
But the most important impact may be that the nursing home industry was ill equipped for the pandemic. The rating system allowed facilities to score high grades without upgrading the care they provided.
“They were working to improve their ratings, but not their quality,” said Charlene Harrington, who sits on a board that advises CMS on the ratings system.
“The problems with the five-star system left these homes less prepared in the pandemic,” she said. “They were allowed to not have enough staffing, and they were allowed to ignore infection-control deficiencies, so they had poorer quality than the public knew about, and they were in the worst position to manage COVID.”
Fulton Commons Care, a nursing home on Long Island, is among hundreds of facilities whose five stars reflected self-reported data that was at times inaccurate.
One of the home’s residents was Stasia Kochanowska, a 75-year-old retired nurse. Last March, her son, John Costa, got a phone call from Fulton Commons: She was being moved from a private room into one with other residents. Costa interpreted that as evidence that the home was short-staffed or overcrowded. “I thought that the idea was to isolate people to keep them safe from the virus,” he said.
Not long after, he got another call: Kochanowska had a fever. On the morning of April 10, Costa got the news he had been dreading: His mother had died from COVID-19.
“Looking back now, I realize that the star doesn’t work and shouldn’t be trusted,” Costa said. “But it’s too late.”
The star system was introduced in 2008. Private equity firms were buying many nursing homes, and lawmakers worried that the quality of care would deteriorate. (Today, for-profit companies own about 70% of all U.S. nursing homes.)
The only way to evaluate nursing homes at the time was a system called Nursing Home Compare. It included tables of statistics about each nursing home. Consumers largely ignored the information.
The new system transformed hard-to-digest information on resident safety, staffing and dozens of other metrics into simple star ratings.
The rating starts with the grade from the in-person inspections and then awards bonus points for facilities that score well on the other two ratings: for staffing and their quality of care. The staffing score reflects the number of hours that employees spend with patients. The quality-of-care rating incorporates 15 metrics — such as residents’ wounds and medications — drawn from what are essentially medical report cards for every resident.
CMS posts each nursing home’s overall star rating, and its ratings for inspections, staffing and quality of care, on a website called Care Compare.
When the system was first rolled out, elder-care groups applauded it for providing an easy-to-understand look inside nursing homes. As it gained popularity, it became a powerful determinant of fortunes in an industry with thin profit margins.
Five-star facilities earned about $2,000 in profits per bed in 2019, according to a Times analysis of nursing homes’ financial statements. Those with three or four stars earned about $1,000 per bed. Poorly rated homes were typically not profitable.
But within months of the star system’s launch, members of a Medicare advisory board cautioned federal officials that nursing homes were incentivized to fudge their numbers, according to board members and former regulators. Board members, including Harrington, a professor emeritus at the nursing school of University of California, San Francisco, called for more auditing of the self-reported data.
Audits, however, remain rare, according to 25 former government officials, congressional aides, nursing home executives and elder-care advocates who meet regularly with CMS officials.
The Times found that from 2017 to 2019 health inspectors wrote up about 5,700 nursing homes, more than one out of every three in the country, for misreporting data about residents’ well-being. That included nearly 800 homes with top ratings. Some facilities didn’t tell the government about certain medications they were giving residents. Others didn’t disclose when residents took serious falls or developed bed sores.
Even when inspectors discovered such misreporting, they didn’t conduct more detailed audits of all of the data that the homes submitted to CMS.
A spokeswoman for CMS said in a statement that the star-rating system is “intended to serve as one tool” for people to choose nursing homes. “Overall scores should be considered in light of a potential resident’s unique care needs,” she said. (CMS would not identify the spokeswoman.)
CMS in recent years “has made multiple substantial improvements to both the ratings methodology and data reliability of the five-star quality rating system. These improvements have made it more difficult to attain or maintain higher ratings, driving providers to continue improving the quality of care they provide,” the spokeswoman said.
She said the Department of Health and Human Services “takes reports of fraud seriously” and that “CMS deploys enforcement and accountability measures swiftly if fraud is detected.”
Not Disclosing Serious Falls
The warnings about self-reported data have been borne out by researchers.
A paper last year by Integra Med Analytics, a data firm started by a University of Texas professor, compared hospital admissions data with the information nursing homes submitted to Medicare. The paper concluded that half of nursing homes underreported potentially deadly pressure ulcers, or bed sores, by at least 50%.
In 2019, the inspector general for the U.S. Department of Health and Human Services found that nursing homes reported only 16% of incidents where residents were hospitalized for “potential abuse and neglect.”
And researchers at the University of Chicago concluded that, from 2011 to 2015, nursing homes didn’t tell Medicare about 40% of residents who were hospitalized after serious falls.
One of the worst offenders in that study was the Clove Lakes home on Staten Island in New York, where 72 residents were admitted to a hospital after a fall, but only 15 were reported to CMS. A Clove Lakes spokeswoman declined to comment.
Menorah Park, a facility in Beachwood, Ohio, that had five stars, disclosed only 11 of 47 serious falls.
Kathleen DeVito, a former resident at Menorah Park, said in an interview that she selected the facility based on its perfect rating. One night, she said, aides took her to the bathroom and then left her alone. She crashed to the floor while trying to wipe herself. Her ankle was shattered, her medical records show. She now has to use a walker.
“I think about how much better I’d be if I had never gone there,” said DeVito, 78, a retired paralegal.
Beth Silver, a spokeswoman for the nursing home, said, “The incidence of falls at Menorah Park is in line with other highly regarded facilities at both the state and federal levels, and are even lower for serious falls, which we attribute to our rigorous fall prevention program.”
She added, “While our sympathy goes out to anyone who may have been injured while a resident at our facility, it does not necessarily mean that the facility was at fault or could have prevented the injury.”
Exaggerating Nurses’ Hours
In 2018, CMS improved the way it collected staffing data. Instead of asking nursing homes to report the average number of hours nurses and aides spent with residents, the agency required homes to submit payroll records that logged their daily hours. The hope was that such data would be harder to manipulate.
With the payroll records, nursing homes’ average staffing numbers plummeted, according to the Times analysis. That suggested that homes had previously been inflating their nursing data.
Even with the new system, many homes still appear to exaggerate how much time nurses spend with patients.
The Times analyzed the publicly available staffing logs submitted to CMS. Thousands of homes — including more than 450 with four or five stars for staffing — derived at least half of their nursing hours from administrators who don’t care for patients. CMS permits the practice, but the public sees only the total number of hours, not the breakdown between administrative nurses and registered nurses who care for patients.
Sun Terrace Healthcare Center had a two-star rating for staffing in early 2019. Then, in the second half of the year, the home, in Sun City Center, Florida, reported a surge in its nursing hours. The bulk of that increase was from the addition of administrators to the tally. By January 2020, Sun Terrace had a five-star rating.
“The administrative nursing staff at Sun Terrace work very closely with the direct care staff in a collaborative team approach to ensure the facility provides the best possible care for each and every resident,” said Alexander Kusmierz, the facility’s executive director.
In light of questions from The Times, Kusmierz said, the home would audit the staffing data it has reported to CMS. “Sun Terrace will continue to make certain that the information provided is thorough and accurate,” he said.
Researchers have determined that the better staffed a facility was, the fewer residents they lost to COVID-19. More employees meant that patients received better care and were more closely monitored. When the pandemic hit, staffing came under additional strain as nurses and workers fell ill.
Trinity Homes, a facility with 230 beds in Minot, North Dakota, received five stars for its staffing levels. But residents repeatedly complained to inspectors that the place was short-staffed. In 2019, half of Trinity’s nursing hours came from administrators.
More than 130 residents there contracted COVID-19. Thirty-five died.
Karim Tripodina, a spokeswoman for Trinity, said the home complies with CMS requirements and has “ample staff to care for our residents.” She said that COVID-19 spread inside the facility at the same time it was spreading in the local community, “despite aggressive efforts to maintain a safe environment.”
‘I Wouldn’t Send My Dog There’
Carrie Johnson, 75, had just had spinal fusion surgery. A raw incision snaked up her back, and it needed to be cleaned regularly to prevent infection. She couldn’t control her bowels.
Using the CMS website, Johnson found a nursing home, Brookdale Richmond Place, where she could recuperate. Part of the country’s largest chain of senior living communities, it was a short drive from her house in Lexington, Kentucky, and the only local facility with a five-star rating for staffing.
“We thought it meant that I’d be safe,” Johnson said.
That was not what it meant.
At the time, in October 2017, Brookdale was operating with a skeleton crew, according to six current and former employees and court documents filed in a lawsuit brought by Johnson. Overwhelmed employees at the 120-bed facility would wake some residents as early as 3:30 a.m. to get them dressed for breakfast and then would put the residents back into bed to wait the hours until mealtime, the lawsuit alleged.
More than a week passed before Johnson was first given a shower, her medical records show. Her surgical wound was not treated. No one came when she rang her call bell for help going to the toilet, so she sat for hours in her own urine and feces. At one point, Johnson said, a passing maintenance man changed her diaper.
On at least three days in her first month there, no one brought her pain medication. Some days there were no nurses on the floor tending to patients, according to the current and former employees and depositions in the pending lawsuit. Nurses spent an average of less than 30 minutes a day per patient.
Brookdale, however, told Medicare that each resident received an average of 75 minutes of daily care. The home arrived at that number by including nurses who were stationed at Brookdale’s corporate headquarters in Tennessee, some 200 miles away, according to depositions and the employees.
Absent the inflated staffing numbers, Brookdale’s rating would most likely have fallen to one or two stars.
Jessica Crutchfield, a nursing assistant who worked at the Brookdale home while Johnson was there, said in a deposition last year that her supervisors told her to falsify residents’ medical records to make it look like they received more care than they did. When she refused, she was fired, she said.
Heather Hunter, a spokeswoman for Brookdale, said, “We have detailed policies in place to ensure compliance with CMS reporting rules, and we are not aware of any instance where inaccurate or false information was submitted by any of our communities outside of the confines of the CMS rules.”
Johnson’s incision began to fester. In November 2017, a physical therapist noticed that it “appeared open and presented with discoloration around incisions.” She alerted the nursing staff, records show, but it took six days before anyone treated the wound.
When Johnson went to an outside doctor, her wound was infected, medical records show. The doctor sent her to the hospital. The infection left her unable to move. Her muscles atrophied.
Hunter didn’t comment on what happened to Johnson. “While The New York Times has the right to print unproven allegations from one side of a lawsuit seeking monetary damages, we choose not to litigate unsubstantiated facts in the press,” she said.
Johnson now lives at home. She used to manage the kitchen at a local homeless shelter; now she can’t walk or even stand. She said she spends about six hours a day in front of the television, waiting for her husband to get back from work. She misses cooking sloppy joes for her six grandchildren.
“I wouldn’t send my dog there,” Johnson said of Brookdale. “That five-star rating is garbage.”
Medicare began tracking nursing homes’ use of certain prescription drugs after the Health and Human Services inspector general found in 2011 that facilities were overusing antipsychotic drugs, which can be especially dangerous for older people. The ratings system now penalizes homes that overmedicate residents.
According to court documents and interviews with nurses, administrators and government officials, nursing homes for years have underreported the number of residents on opiates and antipsychotics.
On Dec. 12, 2017, Eunice Hill checked into a National Healthcare Corp. nursing home in Greenville, South Carolina, for what was supposed to be a short-term rehab stay after a nasty fall made it hard for her to walk. (National Healthcare, a publicly traded company, operates 75 facilities around the United States.) Hill’s daughter, Anne Brown, said she and her mother drew comfort from the facility’s five-star rating.
That first day at the nursing home, staff gave Hill Haldol, a powerful antipsychotic, her medical records show. She was also given anti-anxiety medications every day for a week.
That is not what National Healthcare told the government. At least four times a year, nursing homes are required to provide CMS “minimum data sets” for each patient, the equivalent of medical report cards that detail things like weight, medications and wounds. In Hill’s data set, National Healthcare said that she did not receive any antipsychotics and only got anti-anxiety drugs once.
A National Healthcare spokeswoman declined to comment, citing a pending lawsuit that Hill’s daughter filed against the company.
Hill, 93 and suffering from diabetes, started skipping meals, causing her blood sugar levels to go wild, according to her medical records. In late December, she was rushed to a hospital. Doctors found her kidneys were failing and that she was severely dehydrated.
She died on Jan. 3, less than a month after she arrived at the National Healthcare home.
Rape at Reo Vista
At the Good Samaritan home in Albert Lea, Minnesota, a resident’s foot became infested with maggots in 2018 after her bandages were left on for days. At Dwelling Place in Dover, New Jersey, staff didn’t change residents’ ventilator tubing for more than a month last year, increasing the risk of infection.
In both cases, government health inspectors determined that the nursing homes failed to keep their residents safe. Yet both earned five stars.
The CMS rating system puts the greatest weight on in-person inspections; they are the starting point for the final star rating.
But the exams do little to penalize homes with serious problems. At nearly 1,200 homes with five stars, inspectors over a three-year period identified at least one problem involving potential patient abuse, such as violence committed by staff or the failure to investigate injuries, The Times found.
In 2019, health inspectors concluded that a nursing assistant at the Reo Vista home in San Diego had followed a resident who uses a wheelchair to her room from the kitchen, where she was getting a late-night sandwich. The employee pushed her on the bed and raped her, according to the inspection report.
The report classified the attack as a “category F” violation, a low-level problem that caused potential, not actual, harm.
That designation helped Reo Vista hold onto its perfect rating, which it trumpets on its website.
“This former staff member’s actions do not reflect the values we hold at our facility that our overall five-star rating more accurately reflects,” said William Fredricksen, Reo Vista’s administrator. He said the home was using background checks and other means “to ensure that something like this never again occurs at our facility.”
At 40 other five-star nursing homes, inspectors similarly determined that sexual abuse did not constitute actual harm or put residents in “immediate jeopardy,” The Times found.
After a complaint in 2019, for example, inspectors found that a nursing assistant at Hickory Creek in Madison, Indiana, had barged into the bathroom while a resident was showering and groped her breasts. Hickory Creek kept its five stars.
Hickory Creek officials didn’t respond to requests for comment.
At Manor Care in Boca Raton, Florida, John Schulz, a 57-year-old resident with Alzheimer’s, was strangled to death by another resident in 2019. The state coroner ruled it a homicide. The death doesn’t appear in inspection records at all. The facility has a five-star rating.
“The unfortunate incident was very isolated,” said Julie Beckert, a spokeswoman. She said staff responded “quickly and appropriately” to the strangulation.
In theory, health inspections can occur at any time, without notice.
But The Times found that in 2019 about 70% of nursing homes increased their staff on the days that inspectors visited, compared with their typical staffing levels on that day of the week.
On average, the homes added 25 hours of staff time — a roughly 8% increase — to their rosters on the day of the inspection. For more than 800 facilities, inspection day was the best-staffed day of the year.
The odds of that happening randomly are virtually zero, suggesting that many homes are able to anticipate their inspection dates.
When inspectors visited Mitchell-Hollingsworth Nursing & Rehabilitation Center in Florence, Alabama, on a Thursday in September 2019, the home recorded its highest level of staffing all year: 785 total hours worked by nurses and aides, 15% higher than other Thursdays that year.
“We have absolutely no indication of when the inspectors are coming,” said Brian Scheri, the home’s administrator.
In some cases, corruption was to blame. In 2017, a state health inspector in Florida pleaded guilty to accepting $500,000 in bribes for telling homes about future inspections. Fifteen years earlier, the head of Oklahoma’s health department pleaded guilty to similar charges.
Nursing home staff and state-funded watchdogs said in interviews that hundreds of homes seemed to know in advance about the timing of inspections.
“Every home seemed to get a tip and would roll out the red carpet,” said Tony Cisney, a former inspector responsible for visiting about a dozen homes in Kentucky.
The coronavirus hit America’s nursing homes harder than any other part of the country. Nearly one out of every 10 nursing home residents died from COVID-19.
The pandemic laid bare the flaws in the government rating system.
The state health inspections do little to penalize homes with poor records of preventing and controlling infections. From 2017 to 2019, The Times found, inspectors cited nearly 60% — more than 2,000 — of the country’s five-star facilities at least once for not following basic safety precautions, like regular hand washing. Yet they earned top ratings.
In San Bernardino, California, inspectors wrote up Del Rosa Villa for four different infection-control violations. It kept its five stars. Ninety residents at the 104-bed facility have contracted the coronavirus, and 13 have died.
Del Rosa Villa officials didn’t respond to requests for comment.
Life Care Centers of Kirkland, Washington, the first nursing home in the United States to have documented coronavirus cases, was found in 2019 to have weak infection controls, despite its five stars. State inspectors wrote it up for failing to “consistently implement an effective infection control program.”
Thirty-nine of the facility’s residents have died from COVID-19. The home has 190 beds.
Leigh Atherton, a Life Care spokeswoman, said that citation was the only lapse in infection control that inspectors had identified over 32 previous visits. She said the home quickly fixed the problem.
If the rating system worked as intended, it would have offered clues as to which homes were most likely to have out-of-control outbreaks and which homes would probably muddle through.
That is not what happened.
The Times found that there was little if any correlation between star ratings and how homes fared during the pandemic. At five-star facilities, the death rate from COVID-19 was only half a percentage point lower than at facilities that received lower ratings. And the death rate was slightly lower at two-star facilities than at four-star homes.
A facility’s location, the infection rate of the surrounding community and the race of nursing home residents all were predictors of whether a nursing home would suffer an outbreak. The star ratings didn’t matter.
That is consistent with academic research. Eight recent studies found little relationship between a facility’s star rating and its COVID-19 infection and death rates.
At Fulton Commons Care in East Meadow, New York, its five stars largely stemmed from its self-reported data, which researchers have found at times to be incorrect.
The home reported only 11 instances of residents having serious falls between 2011 and 2015. The University of Chicago researchers, however, identified 40 falls for which residents were hospitalized. Similarly, Integra Med Analytics found that, at most, Fulton Commons was reporting to CMS one-third of its residents’ pressure ulcers. The research firm said the home deserved one star, not five.
Fulton Commons officials did not respond to requests for comment.
Fulton Commons stopped allowing visitors last March. Automated phone calls assured family members that the virus hadn’t infiltrated the home.
Stasia Kochanowska died from COVID-19 on April 10. That same day, another resident, 88-year-old Adele Guyear, spoke to her daughter, Pamela Martin. Usually bubbly, Guyear now sounded terrified. She said her roommate had died from the virus. So had two women down the hall. And, Guyear told her daughter, she had a bad cough.
Martin initially had been comforted by the home’s five-star rating. But during the pandemic she had struggled to get her phone calls returned. Now, hearing from her mother, Martin began to panic.
She said she called the home and demanded that Guyear be sent to a hospital. There, she tested positive for COVID-19. Her mother survived, but she can’t breathe without supplemental oxygen, Martin said.
On May 15, health inspectors cited Fulton Commons for having a deficient program for preventing and controlling infections. CMS classified the problem as causing “minimal harm or the potential for actual harm.” The facility kept its five stars.
Forty-one Fulton Commons residents have died from COVID-19.
“Knowing what I know now,” Martin said, “you’d have to be crazy to give this place a five-star rating.”
The New York Times’ analysis of the star system’s effectiveness was based in large part on data submitted to the Centers for Medicare & Medicaid Services, or CMS. The Times used data from January 2020, before the pandemic upended the industry and temporarily changed data-collection practices. The primary analysis examined the characteristics of homes with five-star ratings, including their staffing patterns and code violations.
The core data set was from the Nursing Home Compare system. That included the most recent set of star ratings for each facility and the underlying metrics. To see how these ratings changed over time, The Times also examined quarterly ratings and metrics going back to 2015.
The Times looked at all code violations reported in the three inspection cycles included in the January 2020 data. The analysis also incorporated nursing homes’ 2019 payroll data (daily logs of the hours worked by nurses, aides and administrators) and Medicare cost reports (financial statements that facilities submit to the government). The cost reports were used to examine the relationship between profitability and the star system.
In addition, The Times used summary data provided by researchers who were granted access to individualized data on Medicare claims. By examining how often nursing home residents were admitted to hospitals after falls or complications from pressure ulcers, the researchers demonstrated the extent to which facilities underreported serious health incidents in data they submit for ratings purposes.