COVID-19 Pandemic Impact on the National Health Care Surveys

Abstract

While underscoring the need for timely, nationally representative data in ambulatory, hospital, and long-term-care settings, the COVID-19 pandemic posed many challenges to traditional methods and mechanisms of data collection.

To continue generating data from health care and long-term-care providers and establishments in the midst of the COVID-19 pandemic, the National Center for Health Statistics had to modify survey operations for several of its provider-based National Health Care Surveys, including quickly adding survey questions that captured the experiences of providing care during the pandemic.

With the aim of providing information that may be useful to other health care data collection systems, this article presents some key challenges that affected data collection activities for these national provider surveys, as well as the measures taken to minimize the disruption in data collection and to optimize the likelihood of disseminating quality data in a timely manner. (Am J Public Health. 2021;111(12):2141–2148. https://doi.org/10.2105/AJPH.2021.306514)

For almost 50 years, the National Center for Health Statistics (NCHS) has collected, analyzed, and disseminated data on health care utilization in the United States. These data have been used to generate nationally representative statistical information and answer key questions about the US health care and long-term-care systems that are of interest to health care policymakers, public health professionals, and health services researchers. To meet its mission, NCHS conducts the National Health Care Surveys (https://www.cdc.gov/nchs/dhcs/index.htm), a family of surveys covering a wide spectrum of health care delivery settings, from ambulatory and outpatient to hospital and long-term-care providers.

While the number of provider and establishment surveys in the National Health Care Surveys family has varied over past decades, there are 5 ongoing data collections as of 2021:

• •National Ambulatory Medical Care Survey (NAMCS): annual survey of office-based physicians and community health centers (CHCs; including federally qualified health centers), which collects practice-level and patient visit data;
• •National Electronic Health Records Survey (NEHRS): annual survey of office-based physicians on their use and experiences with electronic health records (EHRs);
• •National Hospital Ambulatory Medical Care Survey (NHAMCS): annual survey of hospital emergency departments (EDs), which collects organization-level survey and patient visit data;
• •National Hospital Care Survey (NHCS): survey of hospitals providing organization-level and patient visit data from hospitals on inpatient discharge and visits to EDs; and
• •National Post-acute and Long-term Care Study (NPALS; formerly the National Study of Long-Term Care Providers): biennial study designed to provide national and state-level data on multiple long-term-care sectors using survey data on adult day services centers (ADSCs) and residential care communities (RCCs), along with administrative data on home health, hospice, long-term-care hospitals, inpatient rehabilitation facilities, and nursing homes.

These 5 provider- and establishment-based surveys cover a broad spectrum of health care settings and have a combination of design features that make them unique. They are used by policymakers, public health professionals, researchers, and provider and consumer organizations to answer important health care questions and inform policy, research, and practice. These surveys can be used to address health care access and utilization, quality of care, patterns of screening and treatment provided for specific conditions, disparities in health care, and diffusion of pharmaceutical therapies and health care technology. In some instances, these surveys are the only national data source of patient visits made to care settings (e.g., physician offices and CHCs) or that collect characteristics of settings (e.g., ADSCs and RCCs). The majority can be used to analyze trends in health care over time. They have been at the forefront of helping monitor new health information technologies; for example, data from NEHRS are used to monitor interoperability of EHR systems¹ and regulatory or administrative burdens associated with these systems (as detailed in the 21st Century Cures Act).² The absence of fielding these surveys would result in loss of vital data that serve as critical tools for understanding health care delivery and utilization and their impact on public health in the United States.

These surveys are voluntary, and each shares features designed to allow for nationally representative estimates for health care providers. The 2020 and 2021 data collections of the National Health Care Surveys have been significantly challenged by the COVID-19 pandemic, as health and long-term-care providers and establishments continue to serve the nation and battle the pandemic on the front lines. There exists a delicate balance between minimizing data collection burden to these already-overwhelmed care providers and safely collecting health care data that are critically important for monitoring the pandemic and informing health care policies in a timely manner.

This article details strategies that the NCHS has undertaken to mitigate the potential adverse impacts of interrupting NCHS health care survey data collection from across service delivery systems during the COVID-19 pandemic. These strategies allow the NCHS to carry out each of the National Health Care Surveys and add new COVID-19 questions to increase research capacity, while ensuring safety of survey participants, data quality, and timely public dissemination of health care data and statistical information.

AMBULATORY CARE SETTING

NCHS conducts 2 provider-based surveys that collect data on the ambulatory care setting. Both of these were affected by COVID-19.

National Ambulatory Medical Care Survey

Traditionally, NAMCS is conducted in person by US Bureau of Census field representatives (FRs), who administer a computer-assisted personal interview (referred to as the “induction interview”) with a sampled office-based physician, CHC director, or advanced practice provider, followed by manual abstraction of medical records for patient visits made to that physician or provider during a predetermined calendar week.³ In-person methods have been historically used for NAMCS (and NHAMCS) because at the time of the survey’s creation in the 1970s, electronic data systems were not widely available, and in-person collection was considered one of the only manners for collecting detailed health care data uniformly from sampled physicians. The 2020 NAMCS office-based physician component was originally scheduled to begin in March 2020 but needed to be delayed to mid-May because of the initial wave of COVID-19 spreading across the United States (Box 1). Furthermore, because of certain geographic areas being affected differently by COVID-19, real-time data⁴ were used to determine which US counties had (at the time) a high number or rate of positive cases. In these “hotspot” locations, the start of fielding was delayed an additional 4 weeks.

BOX 1 —

Impact of the COVID-19 Pandemic on the National Health Care Surveys: United States, 2021

Timing or Stoppage of Fielding	Changes to Survey Content	Changes to Methods or Modes of Data Collection	Response Rates	Risk and Health Considerations	Changes to Data Dissemination
National Ambulatory Medical Care Survey: Physician Component
Delayed ∼6 wk, with additional 4-wk delay for areas with high COVID-19 counts or rates; visit abstraction ended after 2020 Q1	+5 COVID-19 interview questions (with subquestions) in second half of 2020 and all of 2021	Physician induction interview conducted primarily over phone as opposed to in person; visit data ended after Q1	Preliminary response rates for 2020 physician induction interview: 40.1% in Q1, 49.9% in Q2, 52.9% in Q3, 54.0% in Q4	In-person data collection moved to CATI for physician induction interview	Estimates from the physician induction interview COVID-19 questions released on the NCHS Dashboard
National Ambulatory Medical Care Survey: Community Health Center (CHC) Component
No delays; visit abstraction continued for entire 2020 survey	+5 COVID-19 interview questions (with subquestions) in second half of 2020 and all of 2021	CHC facility and provider induction interview conducted primarily over phone as opposed to in person; no change in collection of visit data	Preliminary response rates for 2020 provider induction interview: 74.0% in first half	In-person data collection moved to CATI for CHC facility and provider induction interviews; remote abstraction available for visit abstraction (if needed)	None
National Electronic Health Records Survey
Fielding delayed from fall and winter 2020 to spring 2021	+7 interview questions on telemedicine in 2020	None	Fielding recently ended; response rate not yet available	No risk or health considerations	None
National Hospital Ambulatory Medical Care Survey
Fielding delayed ∼1 wk, with additional 4-wk delay for areas with high COVID-19 counts or rates	+4 COVID-19 interview questions (with subquestions) in 2021	None	Preliminary response rates for 2020 hospital induction interview: 63.7% in P1, 63.8% in P2, 68.9% in P3 of 2020; for 2020 ED visit abstraction 50.3% in P1, 76.0% in P2, 72.3% in P3	In-person data collection moved to CATI for hospital induction interview; reliance on remote abstraction for ED visit data	None
National Hospital Care Survey (NHCS)
Fielding delayed ∼2 mo	+7 COVID-19 interview questions (with subquestions) in 2020 and 2021	None	Not finalized	No risk or health considerations	Selected preliminary estimates from NHCS 2020 UB-04 claims visit-level data released on the NCHS Dashboard
National Post-acute and Long-term Care Study
Survey component fielding delayed 2 mo from October 2020 to December 2020	+7 COVID-19 questions (with subquestions) to ADCS and RCC questionnaires	None	Not finalized; currently at ∼50% for RCCs and ∼43% for ADSCs	No risk or health considerations	Selected preliminary estimates from the COVID-19 questions released on the NCHS Dashboard

Note. ADSC = adult day services center; CATI = computer-assisted telephone interviewing; ED = emergency department; NCHS = National Center for Health Statistics; P = period; Q = quarter; RCC = residential care communities.

In addition to the delay in fielding NAMCS, there were other challenges. Many physician offices were initially closed or only conducting telemedicine visits, making in-person, on-site abstraction of patient visit data difficult to conduct. Lengthy data abstraction at the sites of care would also pose significant risk for FRs and survey respondents. As a consequence, the administration of the induction interview with sampled providers had to be changed to computer-assisted telephone interviews (CATI). Field representatives for the NAMCS tried to complete manual abstraction of patient visit data for the first quarter of the 2020 NAMCS data collection, which ended in mid-August 2020. The preliminary rate for abstraction for the first quarter of 2020 data collection was 24% compared with 36% in 2019.

As the toll of COVID-19 increased and safety concerns further escalated, in July 2020, NCHS made the decision to continue induction interviews using only CATI and cease all manual abstraction of patient visit data after the first quarter of data collection. The impact of this decision was that there would be no national estimates nor patient visit data available for 2020. While considered, remote abstraction was not deemed feasible because of the technical requirements that needed to be in place at physician offices.

The impact of COVID-19 on the CHC component of the NAMCS differed from the office-based physician component. First, data collection with CHCs was originally scheduled to start in summer 2020; therefore, lessons learned from the NAMCS physician component helped inform the switch to CATI-based induction interviews with CHC directors immediately. Unlike the physician component, as weekly monitoring of the response and abstraction rates occurred in CHCs, NCHS decided that manual abstraction of patient visits to sampled CHC providers would continue on site throughout 2020. This decision was made while coordinating with the US Census Bureau to ensure both intra-agency and federal, state, and local safety measures were being followed to protect both FRs and participants. While preliminary response and abstraction rates were lower than in previous years (i.e., 74% for the first half of 2020 vs 84% for 2019), the rates remained at levels that would generate reliable national estimates, unlike what was seen during the first quarter of visit data collection among physicians.

The practice-level induction interview portion of NAMCS offered a unique opportunity to collect provider-reported data on their clinical experiences during that time. As a consequence, 5 new survey questions about experiences with COVID-19 were added to both the office-based physician and CHC components (Appendix A and Appendix B, respectively, available as supplements to the online version of this article at http://www.ajph.org). These questions asked providers about the following issues:

• •shortages of personal protective equipment;
• •their ability to test patients for COVID-19 or refer them to a testing site;
• •how often patients who tested positive for COVID-19 were turned away from the office or CHC;
• •whether any physicians, providers, or their staff tested positive for COVID-19; and
• •use of telemedicine.

Because of the timing of the creation and approval of these new COVID-19 questions, the items were included in all 2020 NAMCS induction interviews for CHCs but were only fielded in the second half of the 2020 NAMCS office-based physician induction interviews. Furthermore, these questions were scheduled for the entire 2021 NAMCS data collection. To provide more timely data about experiences and impact of COVID-19 on physicians, preliminary results from the new COVID-19 questions were published on the NCHS COVID-19 Data Dashboard⁵ before the release of the final 2020 NAMCS data file.

National Electronic Health Records Survey

NEHRS is conducted through a joint partnership between NCHS and the Office of the National Coordinator for Health Information Technology.⁶ NEHRS is a nationally representative, office-based physician survey about EHR adoption and capabilities, burden associated with EHRs, and progress physicians have made toward meeting the policy goals of the Health Information Technology for Economic and Clinical Health Act. Similar to NAMCS, NEHRS was recognized as a potential data source to collect information about physicians’ clinical experience during COVID-19. New NEHRS questions (Appendix C, available as a supplement to the online version of this article at http://www.ajph.org) were developed to ask about the use of telemedicine in physician offices during 2020 because of the expected increase in telehealth visits⁷ and recognition that more services would be made available through telehealth.⁸

The fielding of 2020 NEHRS was scheduled to occur during fall and winter 2020; however, the start of its fielding was ultimately delayed until March 2021. The reason for this delay was 2-fold: to allow time to revise the survey instrument to include new questions on telemedicine and to ensure that the fielding period avoided November and December, when NEHRS has traditionally yielded lower response rates. Because NEHRS is conducted using a combination of self-administered Web and postal mail responses, there are no safety concerns associated with the administration of this survey. Once NEHRS began, the response rates were monitored on a weekly basis to determine if any adjustments to the fielding schedule or survey operations may be needed.

HOSPITAL-BASED CARE SETTINGS

NCHS conducts 2 surveys that collect data from the hospital-based care setting, both of which were affected by COVID-19.

National Hospital Ambulatory Medical Care Survey

NCHS conducts 2 separate national surveys of hospital-based care settings. The first is NHAMCS, which collects data on the utilization and provision of ambulatory care services in sampled hospital EDs throughout the United States via in-person manual abstraction of ED visits by US Bureau of the Census FRs.³ Fielding of the 2020 NHAMCS was delayed by 1 week until the end of May 2020, close to when NAMCS fielding also began. Similar to NAMCS, NHAMCS fielding was delayed for sampled hospitals located in counties with a high number or rate of positive COVID-19 cases; in these “hotspot” locations, NHAMCS fielding was delayed further by about 4 weeks until June.

Like other National Health Care Surveys, NHAMCS’s traditional approach to conduct in-person induction interviews and manual data abstraction of ED patient visits was not possible because of COVID-19 hospital safety protocols that prohibited FRs from entering the facilities. Therefore, FRs used CATI to administer the NHAMCS induction interview by phone. In addition, NCHS had to change the abstraction of medical records for ED visits from on-site conducted by FRs to various manners of remote abstraction (depending upon the individual infrastructure at each participating hospital), which ranged from hospitals mailing encrypted CDs to US Census regional offices, to submitting records electronically through their secure online portals, and even to having their own staff perform abstraction while being guided over the phone by a FR. Each of these individual methods was vetted by NCHS to ensure strict data security protocols were followed. While limited remote abstraction methods had been established in previous years, the COVID-19 pandemic pushed their expansion. Because remote abstraction methods circumvented COVID-19 restrictions and the sampled hospitals were being extra-accommodating, ED patient visit data for NHAMCS were collected throughout the entire 2020 calendar year. Weekly monitoring of progress showed that while the preliminary response and abstraction rates were lower than in previous years, they continued to remain above 50% and would ultimately generate 2020 national estimates when NHAMCS data collection was completed in May 2021.

Finally, new COVID-19 questions about hospital ED experiences providing care during the COVID-19 pandemic were developed and added to the NHAMCS induction interview with the sampled hospitals (Appendix D, available as a supplement to the online version of this article at http://www.ajph.org). These questions ask about

• •shortages of COVID-19 testing,
• •creation of COVID-19 testing areas outside the ED,
• •ability to test patients for COVID-19, and
• •whether any hospital clinical staff tested positive for COVID-19.

To allow for uninterrupted continuation of 2020 NHAMCS data collection, NCHS decided not to include these COVID-19 questions until the 2021 NHAMCS data collection, which was scheduled to start in spring 2021.

National Hospital Care Survey

The other hospital-based care survey is the NHCS, which leverages electronic health data (primarily EHRs and UB-04 administrative claims) to collect data on patient care in hospital-based inpatient settings and EDs to describe patterns of health care delivery and utilization in the United States.⁹ As with NAMCS and NHAMCS, NHCS collects 2 levels of data: facility and patient visit. But unlike the other 2 surveys, NHCS patient visit data are extracted electronically from health records without needing any in-person manual abstraction. This extraction method remained unchanged during the pandemic because the automated data collection method presented no safety concerns to data collection agents or hospital responders. However, there still was concern regarding increased burden to the sampled hospital participants. The additional care requirements related to COVID-19 have disproportionately increased hospital burden of the provision of care throughout the United States,¹⁰ yet all hospitals are experiencing a strain on their resources and being asked to participate in more data collection and surveillance systems at the federal level than ever before. In recognition of these additional burdens, fielding of the NHCS was delayed by approximately 2 months uniformly across all US regions.

In addition to patient-level data collection, NHCS includes an annual hospital Web-based interview to collect self-reported, hospital-level data about hospital utilization (e.g., number of staffed beds and average length of stay), general information (e.g., primary service type, mergers), and data reporting (e.g., annual visit counts). For the 2020 NHCS hospital interview, new COVID-19 questions were created and added to ask about the sampled hospital’s clinical experiences during the pandemic (Appendix E, available as a supplement to the online version of this article at http://www.ajph.org). These questions cover the following issues related to COVID-19:

• •shortages of testing,
• •creation of testing areas outside the hospital,
• •need to turn patients away to other facilities,
• •whether any hospital clinical staff tested positive, and
• •the number of COVID-19 patients treated during the 2020 calendar year.

Data from the hospital interview have in the past been used primarily for developing weights for the NHCS patient visit data; however, because of the addition of new COVID-19 questions, NCHS is exploring plans to release some of these data to the NCHS Research Data Center (https://www.cdc.gov/rdc/index.htm) for researchers to access.

Given the pressing need for more timely hospital-based data, NCHS began dissemination of rapid, more “real-time” NHCS preliminary patient visit data from inpatient and ED claims for about 50 sampled hospitals. Preliminary unweighted data are released via the NCHS COVID-19 Data Dashboard (https://www.cdc.gov/nchs/covid19/nhcs.htm) and provide estimates on COVID-19 hospital encounters by week, COVID-19 screenings, intubation or ventilator use, mortality, and other related estimates. These preliminary data will serve as another data resource to help inform the monitoring of the COVID-19 pandemic and its impact on hospital utilization.

POSTACUTE AND LONG-TERM-CARE SETTINGS

NPALS is NCHS’ initiative to estimate the supply and use of paid, regulated postacute and long-term-care services providers and their policy-relevant characteristics and practices.¹¹ Administrative data from the Centers for Medicare and Medicaid Services are acquired for home health, nursing home, hospice, inpatient rehabilitation, and long-term-care hospital sectors. Multimode surveys (i.e., Web and mail questionnaires followed by CATI for nonresponse follow-up) are used to collect information from ADSCs and RCCs. Given the disproportionate toll the COVID-19 pandemic has taken on long-term-care sectors including ADSCs and RCCs, operational adjustments in NPALS timing and protocol became essential for the 2020 wave of the survey component.

When COVID-19 started affecting long-term-care providers, questionnaire items were already finalized and ready for the approval process. However, as NPALS is the only federal study that collects national and state-level data on ADSCs and RCCs, new questions were quickly added that ask for

• •the number of COVID-19 cases, hospitalizations, and deaths among ADSC and RCC services users and staff;
• •availability of personal protective equipment and testing kits;
• •changes in visitation procedures;
• •use of telemedicine; and
• •general infection-control policies and practices.

These new questions to the ADSC and RCC questionnaires^12,13 (Appendix F and Appendix G, respectively, available as supplements to the online version of this article at http://www.ajph.org) will allow analysis of challenges faced in long-term-care settings during the pandemic. The process of adding new questions and restructuring tasks (e.g., onboarding and training of field staff) to be conducted virtually shifted the original schedule by 2 months, which necessitated that data collection begin in December 2020. To reduce burden on respondents, telephone calls to verify contact information for target respondents (i.e., the director, administrator, owner, operator) were eliminated.

Over the past several years, response rates for NPALS (as with other health surveys) have declined, and it is expected that response rates may be further affected by the COVID-19 pandemic. As such, existing questions were removed to prioritize COVID-19 questions without increasing the length of NPALS survey instruments. The COVID-19 crisis resulted in several ADSCs closing temporarily and others providing services through alternative methods. To address these scenarios, prospective respondents were advised to complete the survey to the best of their ability, even with services at a center being temporarily or permanently suspended, reduced, or offered through alternative methods, and potentially fewer people receiving services on a regular basis. It is premature to assess response rates, but indications are that the number of ADSCs and RCCs out of business may be higher in 2020 than in earlier years. For instance, a total of 157 ADSCs and 127 RCCs reported as out of business at the end of the 2016 NPALS, whereas, for the 2020 wave, 236 ADSCs and 120 RCCs reported being out of business by the last month of data collection.

For the 2020 wave of the NPALS survey component, before processing of the final data, preliminary data were released on the NCHS COVID-19 Data Dashboard so more timely estimates of the COVID-19 variables could be made available. It is expected that a more stringent evaluation process will be needed to assess final data quality for 2020 relative to earlier NPALS waves, such as examining whether US state differences in response reflect the incidence and impact of COVID-19. On the administrative data side, it is expected that some data may not be “complete” as some surveys of nursing homes have been placed on hold, and requirements to submit assessments during the COVID-19 pandemic may have been waived.

SUMMARY

The COVID-19 pandemic has brought to the forefront the need for reliable, quality, and timely data to monitor the epidemiology of this disease and to inform both national and state policies and guidelines. Nationally representative health care data are particularly valuable as they provide an understanding of health care access and utilization throughout the United States. Collectively, the National Health Care Surveys have been modified and adapted quickly to provide these critical data and help our understanding of care provided in ambulatory, hospital, postacute, and long-term-care settings, as well as other concerns that COVID-19 has brought to light. For example, it has been made clear that social determinants of health (e.g., poverty, physical environment) have a substantial effect on COVID-19 outcomes.^14–16 The National Health Care Surveys result in data that can be used to study these differences: in addition to general care provided for COVID-19, there are patient and services user characteristics and social determinants of health data collected in NAMCS, NHAMCS, NHCS, and NPALS that will allow for study of these inequalities.

When one is using data from these surveys to study health care provided during the COVID-19 pandemic, it is important to understand how the pandemic itself affected data collection. As detailed previously, NCHS needed to adapt traditional data collection strategies for all 5 National Health Care Surveys. All surveys experienced some delays in fielding and a movement to or reliance on remote modes of data collection such as CATI. As such, somewhat lower response rates were expected compared with previous years of data collection. The most notable adjustment made was specific to NAMCS, in that data collection of patient visits in the office-based physician component ended after the first quarter of fielding. As such, there will not be a visit data file available for this component.

Other adjustments NCHS made to the surveys are poised to create a capacity for new research. In each of the surveys, questions were added that will capture the experiences of providing health care during the pandemic. To our knowledge, these questions are not available on other national data sets. They will ultimately be made available to researchers to perform their own data analyses. Furthermore, releases of data from NAMCS, NHCS, and NPALS via the NCHS COVID-19 Data Dashboard marks the first time preliminary estimates were made available from the National Health Care Surveys. While finalized data sets will also be made public or placed in the NCHS Research Data Center, through these preliminary estimates, NCHS is able to provide information to policy- and decision-makers in a timelier manner. Furthermore, it has also created a mechanism that NCHS plans to use in future years—not just for COVID-19, but for other health care topics as well.

While most of these adjustments were to ensure that data collection continued during COVID-19, NCHS expects some to continue on an ongoing basis. The option of using CATI to collect interview data for NAMCS and NHAMCS will continue in the future, in addition to using traditional in-person survey interviews. The use of remote abstraction will continue for NHAMCS as well, allowing hospitals to participate with more flexibility. Finally, while the NCHS COVID-19 Data Dashboard has provided a platform for NAMCS, NHCS, and NPALS to release preliminary estimates for the first time, the exploration of additional mechanisms for releasing preliminary estimates on opioid misuse and abuse, chronic conditions, and other health care topics is also being pursued.

Our experiences conducting the National Health Care Surveys during the COVID-19 pandemic provide several lessons that are important to consider, not only for these surveys but also for other researchers who may themselves be collecting data during the current or future pandemics. The first lesson is to be prepared for potential disruptions and have a willingness to adapt. Each of the National Health Care Surveys, in some way, had their data collections disrupted. However, NCHS adjusted, whether it was to delay or stagger the timing of data collection, to use new modes of survey administration, or even pull certain components of surveys from the field. Ensuring your surveys and data collection systems can make needed adjustments will allow data collection to continue during these challenging times when this collection may be needed more than ever.

Second, topical survey questions or data elements relevant to the study’s population of interest should be included where possible, to collect data on COVID-19 or whatever future crises may arise. This will increase research capacity when one is using the resulting data and ultimately lead to new information that can inform policy and guidelines.

The third lesson is that novel data dissemination methods are also needed. While data collection is critical, disseminating the information to key stakeholders and the public in a timely manner is also important as it allows for informed decisions to be made more rapidly. Finally, the lessons learned from conducting the National Health Care Surveys during the COVID-19 pandemic are not just applicable to NCHS’s current health care data collections but also for future years. Applying the lessons learned from COVID-19 will be important to ensure that NCHS’s National Health Care Surveys are sustainable in the future and continue to be a source of reliable and quality data during any future pandemics. Using these lessons learned will allow researchers not only to continue to collect the data needed for monitoring health care during the COVID-19 pandemic but also to be prepared to better meet the data needs for any future pandemics.