PennLive's investigation, Still Failing the Frail, involved interviews with more than than 40 sources, including experts, families, and elderly care advocates.
It also involved an extensive analysis of documents and data on nursing homes.
Here's a breakdown of the major sources we used, how we used them, and, in some cases, how you can use them to examine nursing homes in your community.
NURSING HOME INSPECTION REPORTS: Nursing home inspection reports were one of most important resources in PennLive's investigation. Inspections are conducted by the department annually and in response to complaints from residents and families. PennLive read over hundreds of pages of reports to gauge recurrent and serious citations in Golden Living's homes. To see inspections in a Pennsylvania nursing home you can consult either the state Department of Health website or you can find them by searching for specific nursing homes on Nursing Home Compare, the federal government's nursing home comparison website.
NURSING HOME CITATION DATA: While inspection reports provide the details behind why a home was cited, PennLive also wanted to examine the raw number of citations issued against homes. Those numbers are tracked by the federal government on its consumer website, Nursing Home Compare. You can view citations for nursing homes across the country for the past three years here and archived data here.
NURSING HOME PENALTIES: When the state Department of Health determines a nursing home has committed a serious violation, it can do one of three things: recommend the federal government issue a penalty, issue a penalty itself, or do both. In its 2016 investigation, PennLive found that the Pennsylvania Department of Health was routinely failing to take any punitive action against serious violators. To analyze how often the department penalized Golden Living's former homes, PennLive used two sources: federal penalties data on Nursing Home Compare and a list of state penalties on the Department of Health's website.
STAFFING DATA: One of the major ways experts assess nursing home quality is staffing levels. Research has consistently shown that high nurse-to-resident ratios are one of the most important determinants of good care. To compare staffing levels between Golden Living and its new owners, PennLive turned to two different sources: For staffing data prior to 2017, PennLive analyzed Medicaid cost reports. These reports provide detailed information on the number of hours worked by different types of nursing staff. For more recent staffing information, PennLive used data from a new federal data collection system called the 'Payroll Based Journal.' PennLive consulted Charlene Harrington, professor emeritus of nursing and sociology at the University of California, San Francisco, while analyzing this data. Because of concerns about the accuracy of PBJ data in the first half of 2017, PennLive only used data from the second half.
NURSING HOME LICENSE APPLICATIONS: To better understand what kind of scrutiny the Department of Health gave to the nursing home chains that bought Golden Living's nursing home licenses, PennLive read through hundreds of pages of nursing home licensing documents. Those documents are not available online and were acquired by PennLive through a request under Pennsylvania's open records law. The department withheld some records, arguing they were confidential and proprietary. PennLive appealed to the state Office of Open Records for the release of these records. At the time of writing, the office still hasn't reached a determination. You can file your own request for nursing home licensing information by following these instructions and addressing your request to the Pennsylvania Department of Health.
OWNERSHIP INFORMATION: Because for-profit nursing chains are typically structured in complex ways, it can be difficult to understand who runs a company. Fortunately, the federal government requires nursing homes to provide certain ownership information. To determine who was running the companies now operating Golden Living's former homes, PennLive used ownership data stored here. The information was not always perfect or up-to-date, but it was the best available.
REAL ESTATE INFORMATION: Somewhat unusually, Golden Living retained ownership of the real estate of all 36 of its nursing homes when it transferred its licenses to new operators. To determine whether Golden Living continued to retain the real estate to those homes, PennLive scoured deed and assessment reports collected by LexisNexis, a subscription service that collates public records. While LexisNexis is a paid service, that same information is largely available for free on most county websites. For instance, to find who owns the property to a nursing home in Dauphin County you can search property records here.
FINANCIAL DATA: Because most for-profit nursing home chains are private companies, it was difficult for PennLive to determine the profit margins of the new operators of Golden Living's homes. PennLive was able to glean some financial information by analyzing financial documents - called "cost reports" - that nursing homes are required to file with state and federal governments. The state Department of Human Services posts Medicaid cost reports on its website here. You can download PDFs of cost reports for specific homes by clicking on 'MA-11 Resources' and 'Accepted Cost Reports'. These documents were also essential for PennLive's analysis of 'related party transactions'. Those are transactions that, advocates and some experts say, are used by nursing homes to disguise their profit margins.
MEDICAL RECORDS: As part of its series, PennLive spoke with Richard and Ashley Mantheiy, a couple that believe a nursing home's poor care led to the death of Richard's mother. To vet their story, PennLive reviewed hundreds of pages of medical records produced by the nursing home and by medical professionals that treated Richard's mother.
OTHER SOURCES: To gain a better understanding of the companies now running Golden Living's former nursing homes, PennLive also turned to an array of other sources. It used company filings in Pennsylvania and other states to understand connections between nursing home investors. It looked at UCC filings to understand how often those companies were getting loans and from who. In some cases, PennLive used court records to provide additional details about those owners.
About our nursing home databases
As part of its reporting on nursing homes, PennLive has posted two databases:
- Golden Living database:
- Pennsylvania database:
The information is largely drawn from Medicaid cost reports filed with the Pennsylvania Department of Human Services. PennLive has cleaned, compiled and, in some cases, analyzed that data, to make it easier for the public to search and understand.
About our analysis
The raw data for the Medicaid cost reports PennLive analyzed can be found on the Pennsylvania Department of Health's website here: http://nfrp.panfsubmit.com/
Although much of PennLive's data is drawn directly from these reports, it's worth noting that some fields involved calculations by PennLive.
For instance, PennLive converted raw hours worked by different types of direct care staff to a measurement known as 'hours per resident per day' or 'HPRD'. This is a measurement of how many hours, on average, different types of direct care staff spent with residents each day.
For certified nursing assistants, this is what that calculation looks like:
Total hours worked by CNAs / number of resident days = CNA HPRD
This metric is used by researchers and the federal government to compare staffing levels between nursing homes. Researchers and advocates recommend facilities provide 4.1 HPRD of direct care staffing (meaning certified nursing assistants, licensed nurse practitioners and registered nurses).
It's worth noting that Pennsylvania requires nursing homes to only provide at least 2.7 HPRD. For this reason, elderly care advocates believe the state needs to increase this threshold.
Because 'HPRD' can be difficult for readers to understand, PennLive also created another field in its database: 'Staffing Level' (also referred to as 'PennLive Staffing Score' in the databases).
This field categorized a home's staffing as either 'Dangerously low', 'Very low', 'Low', 'Good', or 'Very good' based on its average hours per resident day.
PennLive partially determined those categories based on the 4.1 HPRD recommendation and a research paper on recommended minimum staffing standards written by several leading nursing home researchers. PennLive also consulted one of the paper's authors, Charlene Harrington, professor emeritus of sociology and nursing at the University of California, San Francisco.
For instance, PennLive defined a facility with 'dangerously low' staffing as one that provided a less than 3.5 HPRD total direct care staffing and, within that, provided less than 0.53 HPRD from registered nurses.
You can read more about those categories and other fields in PennLive's databases at the end of this article.
Tips for analyzing our data
If you're a researcher, reporter, elderly care advocate, nursing home employee, or simply a member of the public who wants to know more about a particular nursing home, there's a lot you can learn about Pennsylvania facilities from our data that isn't available on state or federal websites
Here's some suggestions of things to look at:
- Pay:
- Administrator Pay:
- Nursing home profit/loss
- Related party transactions:
- Home office expense:
- Staffing:
- Rent:
If you're a researcher, elderly care advocate, or a reporter who wishes to use PennLive's data, please feel free to do so. All we ask is that you provide some attribution to PennLive.com.
If you have further questions about PennLive's analysis or our methodology, please contact reporter Daniel Simmons-Ritchie at daniel.simmons.ritchie@pennlive.com
More information about each field in our statewide database:
Report_year
The fiscal filing year for a given cost report.
Facility
Nursing home name as drawn from cost report.
City, Address, Zip_code, and State
Because the above fields are not included in Pennsylvania's downloadable cost report files, PennLive added these fields by connecting cost reports with datasets maintained on the federal government's Nursing Home Compare website. Address information is unavailable for some homes.
Medicaid_report_number
This is the unique identification number of a cost report.
Report_period
The period that a cost report encompasses.
Facility_medicaid_number
Each nursing home has a unique Medicaid number. PennLive has included this field for analysts who wish to connect PennLive's dataset with government datasets.
Federal_provider_number
The facility's federal provider number. This can be used to join this dataset with Nursing Home Compare datasets. Note that this information may be missing for some homes.
Home_office
Nursing homes that are in a chain typically have a 'home office' or headquarters. Records that note, 'see organizational chart', are a note from the nursing home to the state Medicaid program.
Management_company
Some nursing homes contract operations to management companies.
Other_controlling_entity
A field for nursing homes to indicate whether they are controlled by a different entity than those listed above.
Organization_type
Whether a home is a for-profit, non-profit or governmental institution.
Bed_capacity
Number of beds in a nursing home.
Percent_of_residents_on_medicaid
The percent of nursing home residents on Medicaid. This figure is important because nursing homes are reimbursed at a far lower rate for Medicaid residents than Medicare residents. The higher a nursing home's Medicaid population, the less revenue it likely has.
Amount_spent_on_rent
The amount of money a nursing home spent on rent over the cost reporting period.
Total_capital_costs
The amount of money a nursing home spent on capital costs over the cost reporting period.
Gross_revenue
A nursing home's gross revenue over the cost reporting period.
Total_expenditure
A nursing home's total expenditure over the cost reporting period.
Net_revenue
A nursing home's net revenue.
Profit/loss
A nursing home's net income, aka profit/loss, over the cost reporting period.
Profit/loss margin
A field calculated by PennLive by dividing net_income into net_revenue.
Total_spent_on_related_business_parties
Total spent on transactions with related business entities.
Amount_sent_to_home_office
This figure represents part of a nursing home's 'total spent on related business parties'. It represents the amount of money a home sent to its 'home office' or headquarters. These payments are intended as compensation for services, eg. accounting or HR, provided by an organization's home office.
Related_party_1_name
If available, the name of the the first related party a home has listed in its cost report.
Related_party_1_amount_spent
If available, the amount spent by the home on the first related party a home has listed in its cost report.
Related_party_2_name
If available, the name of the the second related party a home has listed in its cost report.
Related_party_2_amount_spent
If available, the amount spent by the home on the second related party a home has listed in its cost report.
Related_party_3_name
If available, the name of the the third related party a home has listed in its cost report.
Related_party_3_amount_spent
If available, the amount spent by the home on the third related party a home has listed in its cost report.
Average_hours_of_care_per_day_per_resident_by_cnas
Calculated by PennLive by adding total hours worked by contract and non-contract CNAs and dividing by 'actual resident days for period' in cost report. Values of >12 are eliminated as potential data errors.
Average_hours_of_care_per_day_per_resident_by_lpns
Calculated by PennLive by adding total hours worked by contract and non-contract LPNs and dividing by 'actual resident days for period' in cost report. Values of >12 are eliminated as potential data errors.
Average_hours_of_care_per_day_per_resident_by_rns
Calculated by PennLive by adding total hours worked by contract and non-contract RNs and dividing by 'actual resident days for period' in cost report. Values of >12 are eliminated as potential data errors.
Average_hours_of_care_per_day_per_resident_by_all_direct_care_staff
Total average staffing hours per day was calculated by adding CNA, LPN, and RN HPRD together. Values of <1 or>12 are eliminated as potential data errors.
Pennlive_staffing_score
A grading system PennLive created to help readers understand how well a home is staffed. The grading system works as follows:
If total staffing hours <=3.5 and RN staffing <0.53: "Dangerously low"
If total staffing hours <= 3.5: "Very low"
If total staffing hours <=4.1 and>3.5: "Low"
If total staffing hours <=5 and>4.1: "Good"
If total staffing hours >5 and <12: "Very good"
If total staffing hours was more than 24 or CNA/LPN/RN hours were more than 12 then the field was left blank because it's likely there's an error in the data.
Administrator_name
The name of the nursing home administrator at the time the cost report was filed.
Administrator_pay
The amount paid to an administrator or administrators over the course of the cost reporting period, excluding benefits. Note that because an administrator may have changed during the reporting year it's possible the listed administrator received only part of this figure.
CNA_wage_per_hour
PennLive calculated this value by dividing the total a home spent on non-contracted CNAs by the total number of hours worked by non-contract CNAs. Note: small numbers of hours worked may create the appearance of very high wages per hour. Extreme values (those higher than $200 per hour) are assumed to be data errors and have been eliminated.
LPN_wage_per_hour
PennLive calculated this value by dividing the total a home spent on non-contract LPNs by the total number of hours worked by non-contract LPNs. Note: small numbers of hours worked may create the appearance of very high wages per hour. Extreme values (those higher than $200 per hour) are assumed to be data errors and have been eliminated.
RN_wage_per_hour
PennLive calculated this value by dividing the total a home spent on non-contract CNAs by the total number of hours worked by non-contract RNs. Note: small numbers of hours worked may create the appearance of very high wages per hour. Extreme values (those higher than $200 per hour) are assumed to be data errors and have been eliminated.
Contract_CNA_wage_per_hour
PennLive calculated this value by dividing the total a home spent on contract CNAs by the total number of hours worked by non-contract CNAs. Note: small numbers of hours worked may create the appearance of very high wages per hour. Extreme values (those higher than $200 per hour) are assumed to be data errors and have been eliminated. This field may also be blank if a home does not use contract staff.
Contract_lpn_wage_per_hour
PennLive calculated this value by dividing the total a home spent on contract LPNs by the total number of hours worked by non-contract LPNs .Note: small numbers of hours worked may create the appearance of very high wages per hour. Extreme values (those higher than $200 per hour) are assumed to be data errors and have been eliminated. This field may also be blank if a home does not use contract staff.
Contract_rn_wage_per_hour
PennLive calculated this value by dividing the total a home spent on contract RNs by the total number of hours worked by non-contract RNs. Note: small numbers of hours worked may create the appearance of very high wages per hour. Extreme values (those higher than $200 per hour) are assumed to be data errors and have been eliminated. This field may also be blank if a home does not use contract s
Ad Choices