Frequently Asked Questions

Learn more about our survey questionnaires, accessing NDWS data, the available data linkages, and more.

If you have additional questions after reviewing the FAQs, email us at info@ndws.org.

NDWS General Questions

Who has access to NDWS data?

NDWS data are available to all researchers at no cost.

Are there any charges for accessing NDWS survey data or the available linked data, including CMS claims data?

All NDWS data, including data linkages listed on our website, such as CMS data, are free to researchers.

Survey Questionnaires/Sample Frame

Is the sample design nationally representative?

The goal is for the sample design to be nationally representative. In Wave 1, the sample is truly national for the Community Clinician survey and the Nursing Home survey. For the Wave 1 Assisted Living and Home Care surveys, the sample frame represents assisted living communities and home health agencies from about 20 states. Note: The Wave 1 survey data release won’t contain data from the Home Care Staff survey. The Wave 2 samples for all four surveys are truly national.

What, if any, steps are being taken to validate survey responses to ensure internal consistency/accuracy?

Surveys include built-in validation checks to ensure internal consistency and accuracy. These checks validate consistency between responses (e.g., total beds used ≤ total beds available), apply date and range checks (e.g., residency year vs. degree completion year), and enforce mutually exclusive multiple-choice options. Hard checks confirm demographic data, such as the year moved to the U.S., as well as soft checks, for example if respondents indicated they did not currently have a clinical job. We also evaluate paradata (e.g., section timings), pre-fill some questions (from respondents' pre-survey screening or sample data), and follow up on incomplete responses to improve data quality.

Have any of the survey questions been used in previous surveys? If so, are data from those surveys publicly available?

Many of the questions were adapted from other sources rather than being exact. Download the Survey Source Crosswalk on the NDWS website. Each question's source (and link) is in the farthest right column. If there is no source, it means the question was developed by the NDWS survey team. The available links can be used to investigate if data from previous studies are publicly available.

What types of staff are you collecting data from? Did you collect data from social services staff or social workers in Wave 1?

Social workers and social services staff aren’t included in Wave 1, though they may be added in future waves. Job titles included in Wave 1 are below.

Community Clinician	Nursing Home	Assisted Living	Home Care
Physicians in Internal and Family Medicine (including Geriatrics, and Hospice and Palliative Care), Psychiatry, and Neurology	Registered Nurse	Health Care Supervisor	Registered Nurse
Nurse Practitioners in Primary Care and Psychiatry	LPN/LVN	Registered Nurse	LPN/LVN
Non-Surgical Physician Assistants	CNA	LPN/LVN	CNA
	Nurse Aide in Training	Direct Care Worker	Nurse Aide/Assistant
	Medication Aide/Technician	Nurse Aide/Assistant	Home Health Aide/Assistant
		Personal Care Aide/Assistant	Personal Care Aide/Assistant
		Medical Care Technician/ Medication Aide	Activity Staff
		Activity Staff

Are you sampling organizations affiliated with the VA?

VA facilities are not part of our sampling plan at this time.

Do the sample sizes allow for subgroup analysis?

It depends on the subgroup. The Community Clinician survey has the largest Wave 1 sample size and thus will allow more possible subgroup analyses. However, the other surveys have smaller samples in Wave 1, so subgroup analyses will not likely be feasible.

Is there a variable that we can use to group clinician respondents within clinics? How many clinicians on average are surveyed from the same clinic?

No, CC respondents can’t be grouped within a specific clinic or practice setting. Given the national sample, which was not created at the organizational level (unlike the staff surveys), it is extremely unlikely that two clinicians from the same clinic would have been sampled.

Accessing the Data

On the NIA LINKAGE platform, do researchers have to be familiar with SAS to use the NDWS survey data and linked data? Can we use other software like R within the platform?

Per the NIA Data LINKAGE webpage: "This platform offers a secure, easy-to-navigate, and remotely accessible statistical analysis environment that meets federal data security requirements for protecting personally identifiable and private health information. Users have access to SAS, STATA, R, Python, MS Office applications and more."

Do researchers need to have a current NIH-funded grant in order to use the restricted data available on the LINKAGE platform or apply for pilot funding?

No, a current or prior record of NIH funding is not needed to access NDWS restricted data via the LINKAGE platform or to apply for pilot funding.

If the linked data I want to use with the Community Clinician survey are included in the Claims-Based Summary File on your website, do I need to request additional linkage through a separate DUA?

Because any/all of the linked data are only available through the LINKAGE program, researchers still need to complete a DUA for access. However, projects using linked data that are not CMS beneficiary-level data have a simpler DUA request process than those using the beneficiary-level data.

Can a research assistant (undergraduate student) receive access to the LINKAGE platform? If so, are there any associated fees?

As part of an approved project, researchers can identify the individuals who should have access to their designated workspace in LINKAGE (e.g., a research assistant or analyst). There are no fees for use of LINKAGE, regardless of the specific data used or the number of approved users on a given project.

Does NDWS provide the state where the participant is residing?

The state of survey respondents is available in the restricted-access version of the surveys available on LINKAGE.

Data Linkages - Linking NDWS Surveys

Does the NDWS data support linking between the administrator and staff survey data sources? And, are they linked in the public-use files?

Yes, the NDWS data supports linking between the administrator and staff survey data sources. No, these aren’t linked in the PUF. If you want to use any information from the administrator surveys as part of your study of staff, you will need to request linked data and conduct your analyses in the LINKAGE platform. There is no charge for access through this route.

Is it possible to link a community clinician to the nursing homes/assisted living facilities where they provide care?

NDWS survey data don’t include direct linkages between clinician NPI and the nursing home CCNs where they practice, if they provide care in a residential setting. The researcher would need to request survey data linked with beneficiary-level CMS data via LINKAGE and follow established methods to link clinicians with facilities using the beneficiary-level data obtained through LINKAGE. At this time there are no identifiers in CMS data that uniquely identify assisted living communities.

Data Linkages - CMS Data

When linking surveys to Medicare claims data, is the linkage limited to only the visits where the clinician surveyed is providing the services? If the surveyed provider says they refer to neurologists and I want to examine if their patients saw neurologists, would this be possible? Similarly, would I be able to examine records for these patients in prior years to establish incidence of dementia diagnosis?

CMS data are available for all beneficiaries that a survey respondent saw, including their encounters and assessments with other clinicians (or in other nursing facilities). Data are available from 2015, so researchers should be able to identify when dementia was first diagnosed (provided it was not before 2015).

Can beneficiary-level claims data be linked to individual nursing homes for which NDWS staff survey data are available, via the NIA LINKAGE program?

Yes. For the Nursing Home Staff survey, NDWS surveys in LINKAGE include the CCN, which can then be used by LINKAGE to identify patients who received care in those participating facilities (through the MDS assessments conducted in that facility). Then, with those patients, LINKAGE can connect to Medicare and Medicaid beneficiary-level claims and encounter information. The summary files referred to on the LINKAGE website are things like the MBSF Chronic Conditions file and Cost & Utilization file. (One additional thing to keep in mind is that there is a lag in the availability of CMS data to researchers, so CMS data available will not be from the same time as survey fielding.) Researchers will not have access to identifying employer information to do their own linkages. Please note that beneficiary-level claims can’t be linked to individual assisted living communities or home health agencies with survey data in Wave 1, but we aim to have this linkage available in subsequent years.

Would potentially linked Medicare data include beneficiaries enrolled in Medicare Advantage, in addition to FFS enrollees? If yes, what years of fee-for-service claims data and Medicare Advantage encounter data are available for linkage?

Researchers will have access to--with appropriate justification--any CMS data available in NIA LINKAGE, which includes traditional Medicare (1991-2023), Medicare Advantage (2015-2021), and Medicaid (MAX from 1999-2015; TMSIS from 2014-2022) data. Because of the lag in CMS data availability for research projects, any and all CMS data requested will be from years prior to survey fielding.

Can we access Medicare Fee-For-Service (FFS) claims data for beneficiaries without a documented dementia diagnosis to assess screening rates among this population?

Yes, the linked Medicare claims include all beneficiaries a clinician has seen, not just those with a dementia diagnosis.

When is the 2024 FFS data expected to be available for linkage? Currently, the NDWS website lists Medicare Advantage claims data availability from 2015-2021. Will the 2022 or 2023 Medicare Advantage data be accessible, and if so, are there anticipated delays in obtaining these datasets?

CMS data is generally available through LINKAGE at the same time as it is available to individual researchers through the VRDC. Projects using LINKAGE (e.g., NDWS or NHATS) request new files through an updated DUA; these data are then available to individual research teams. After the initial Wave 1 data release, NDWS plans to update its DUA for CMS data on a regular, annual basis. Each update will include the new CMS data that have become available in the interim. At this time, we tentatively plan to update our DUA by November 2025; that DUA update will include whatever additional CMS data have become available to projects in LINKAGE by that time. In the experience of the NDWS team, we generally anticipate new CMS data to be available about 12 months after the last release of that specific file (dates can be found through ResDAC). This can help anticipate what additional files might be available by Fall 2025 (e.g., the most recent year of MA data that might be expected by then) when we update our DUA. However, with recent changes in the federal government, the anticipated timing of data release may change significantly.

Will it be possible to link the agency-level CMS Care Compare measures to the Home Care Staff survey, and further link both to geographic information from the Administrator survey?

Please note that Home Care Staff survey data and this linkage won’t be available in Wave 1. In future waves, Care Compare Home Health Quality Reporting will be available for linkage with this survey. Researchers will not be able to upload any additional files to LINKAGE; however, all NDWS surveys may be linked to the specific geography-related sources that are outlined on the Data Linkages page. These linked data will be prepared for researchers by the LINKAGE and NDWS teams, but the specific geographic information collected in the surveys won’t be available to researchers (i.e., researchers will have a file prepared where a Home Care Staff survey respondent from an organization in Michigan would be linked to the appropriate Michigan data from the AARP LTSS Scorecard, but then the "Michigan" state value would be replaced by an encrypted value for state [or county, if using AHRF or SDI] in the file prepared for the researcher).

Currently, the claims-based summary files are based on Medicare data from 2022. Will Wave 1 data users be able to access summary files based on 2024 data?

The claims-based summary files will not be updated until the Wave 2 data release that will occur in 2026. (Given the lag of CMS data availability, the claims-based files will always lag several years behind the period of survey fielding.)

In addition to the claims-based summary file that focuses on aggregated clinical characteristics and outcomes of persons living with dementia cared for by NDWS participants, is it possible to pursue linkage to additional beneficiary-level claims data (including CCW files) by linking NDWS survey respondents to claims-level, rendering/attending provider NPI?

Yes. As outlined on the Data Linkages webpage, through the LINKAGE platform researchers can use respondent encrypted NPI (for the Community Clinician Survey) or CCN (for employing organization of respondents to the Nursing Home Staff Survey) to link to the following beneficiary-level CMS data: traditional Medicare, Medicare Advantage, Medicare Part D, Medicaid (MAX and TMSIS), Minimum Data Set, and OASIS. More recent CMS data will be made available to researchers as it becomes available within LINKAGE. Note: CCN will be also available for employing agencies of respondents to the Home Care Staff survey in future waves when Home Care Staff survey data is available.

Would it be possible to obtain unique CMS Certification Numbers (CCN) for each of the facilities involved?

No, researchers will not be able to obtain the unique CCN for the participating facilities. Unique, encrypted identifiers will be used for the purposes of the available data linkages as outlined on the Data Linkages webpage, but researchers will not have access to the actual CCNs.

Data Linkages - Other Sources Available for Linkage

Can I provide input on additional public datasets to link to the staff and administrator surveys?

Wave 1 sources are set, but we welcome recommendations and suggestions for additional sources we could make available in LINKAGE in future years. Email us at info@ndws.org.

For the administrator and staff surveys, do the public use files or the datasets available on NIA LINKAGE contain identifying employer information (e.g. provider number, name, address) to allow researchers to link to external data sources?

Neither the public use files nor the datasets available via NIA LINKAGE contain identifiable employer information that will allow researchers to do their own linkages to external data sources. For the datasets available via LINKAGE, researchers may request to have the survey data linked at the organizational level by NIA LINKAGE personnel to data sources including Nursing Home Compare, Payroll Based Journal Data, LTCFocus, Area Health Resources Files, CMS Provider of Service File, and a series of NDWS-generated organizational measures. The NDWS-generated measures are available for nursing homes only in Wave 1 but will be available for assisted living communities & home health agencies in future waves. The data dictionary for these NH measures is available on the Data Linkages webpage. Researchers are not currently able to import other external data not housed on NIA LINKAGE for linkage to survey data.

We are interested in exploring trends in dementia care delivery in rural areas and in proximity to Indigenous communities. Is it possible to access geocoded information via NIA LINKAGE? If not, are RUCA codes for surveyed organizations available for analysis?

These are available for the Assisted Living Frame in Wave 1 and will be for Home Care in future waves. We have the raw RUCA codes (1-10) and a recoded version where RUCA codes 7-10 are coded rural=1. We will include the geo-coded RUCA variable on the LINKAGE files for the Community Clinician and Nursing Home sample frames.

Could you clarify what geographic scale (ZIP code, Census Tract) the geo-coded RUCA code was derived from and whether the raw RUCA code will also be included in the file? Also, are Health Professional Shortage Areas (HPSAs) and Medically Underserved Areas (MUAs) indicators included within the provided linkages? If not, will the provided data allow using either Provider or Facility codes, or indirectly using information from AHRF?

The RUCA codes in the NH and CC surveys were assigned by geocoding to the Census Tract. For the AL survey, we connected the RUCA code via the ZIP code. These RUCA codes (not just dichotomous urban/rural) will be available in LINKAGE. The full AHRF will be available and can be linked to surveys using an encrypted--not actual--FIPS county code. No version of the survey data (PUF or available in LINKAGE) will include unencrypted clinician, facility, or geographic identifiers.

Does the de-identified, public-use NDWS survey data include the state variable? If geographical data is not available in the public dataset, would it fall under restricted data that researchers can apply for?

The public-use surveys do not include geographic variables. At this time, the ability to link to geographic information can only be done on the LINKAGE platform and is limited to the specific data sources outlined on the website under the "Environment" section of the Data Linkages webpage. Researchers will not be able to bring their own data sources to LINKAGE, though we hope to expand the linked data available in the future.

Are the total number of staff accessible as part of this dataset? If not, is it possible to connect the Nursing Home Staff Survey with the AARP Long-Term Services and Supports Scorecard to calculate staff turnover rate? It appears that these two datasets would need to be linked by geography (state or county level). Does this mean that I can only obtain turnover rates at the state or county level, rather than at the facility level?

Researchers will have access to the data sources listed on the Data Linkages webpage, and any variables that they contain. In other words, any variable available in Nursing Home Care Compare and Provider Data Catalog Consolidated Data Dictionary will be available in the version of the file available on LINKAGE (with the exception of the actual CCN or other facility-identifying information). In addition to linking the NDWS survey data with NH Care Compare, researchers can also request to link to the state-level AARP LTSS Scorecard. Finally, turnover could also be calculated using the Payroll Based Journal Data that will be available for each participating facility.

For the available linkages, is it possible to link the NDWS data with the Health and Retirement Study (HRS) or other NIA-funded surveys?

No. The only data available for linkage are listed on the Data Linkages webpage.