Catharine W. Burt

Understanding and Interpreting the National Hospital Ambulatory Medical Care Survey: Key Questions and Answers

SEE EDITORIAL, P. 722. Editor’s Note: The National Hospital Ambulatory Medical Care Survey (NHAMCS) is widely used for medical research. Nearly 500 articles have been published based on this database, including 28 in Annals of Emergency Medicine. NHAMCS is a national probability sample survey of visits to emergency and outpatient departments in nonfederal, general, and short-stay hospitals, conducted by the Centers for Disease Control and Prevention, National Center for Health Statistics (http://www.cdc.gov/nchs/ahcd.htm). Strengths of the NHAMCS survey are its rigorous methodology, nationally representative nature, large size, wide array of variables, and capacity to examine long-term trends. Investigators can freely download the database and test locally developed hypotheses. Challenges with NHAMCS are that, given its logistic and statistical complexity, resulting research can be difficult for readers to interpret and for editors and reviewers to critique. There are important limitations and caveats to this survey that, if overlooked, could lead to misleading or inaccurate conclusions. In this article, we have a posed a series of questions to 2 NHAMCS statisticians, and we believe that their answers will be invaluable to both NHAMCS investigators and readers of its research.

Summary statistics for acute cardiac ischemia and chest pain visits to United States EDs, 1995–1996

In this article we describe characteristics of emergency encounters for patients with a diagnosis of acute cardiac ischemia (ACI) and for patients with chest pain complaints. Data are from the National Hospital Ambulatory Medical Care Survey, which includes abstracts from the medical records of a national probability sample of visits to emergency departments (ED). Analysis was limited to records of patients 25 years of age and older with a diagnosis of either confirmed or suspected acute myocardial infarction (AMI) or unstable angina pectoris and records with a nontraumatic chest pain complaint. There was an estimated annual average of 1.2 million visits to EDs by patients 25 years and over with a diagnosis of ACI in 1995-1996, an average annual rate of 7.2 visits per 1,000 persons. Visit rates varied by patients age, race, and gender. Chest pain was a complaint in three-fourths of all ACI visits. There were an estimated 4.6 million annual ED visits where in patients aged 25 years and older had complaints of nontraumatic chest pain, an average annual rate of 27.7 visits per 1,000 persons. ACI accounted for 11% of all chest pain visits, but the probability of the chest pain visit having an ACI diagnosis varied by patients age and race. There remains a large amount of variation in treatment for suspected and confirmed AMI, and for patients presenting with chest pain to EDs.

Guide to Using Masked Design Variables to Estimate Standard Errors in Public Use Files of the National Ambulatory Medical Care Survey and the National Hospital Ambulatory Medical Care Survey

Until recently, sample design information needed to correctly estimate standard errors from the National Ambulatory Medical Care Survey (NAMCS) and the National Hospital Ambulatory Medical Care Survey (NHAMCS) public use files was not released for confidentiality reasons. In 2002, masked sample design variables were released for the first time with the 1995–2000 NAMCS and NHAMCS public use files. This paper shows how to use masked design variables to compute standard errors in three software applications. It also discusses when masking overstates or understates “in-house” standard errors, and how masking affects the significance levels of point estimates and logistic regression parameters.

Are There Patient Disparities When Electronic Health Records Are Adopted

Using nationally representative samples of visits from the 2005–2006 National Ambulatory Medical Care Surveys and the National Hospital Ambulatory Medical Care Surveys (N=39,343), this study examines whether electronic health record (EHR) systems have been adopted by primary care physicians or providers (PCPs) for poor minority patients at the same rate as by the PCPs for wealthier non-minority patients. Although we found that electronic health record adoption rates varied primarily by type of practice of the PCP, we also found that uninsured Black and Hispanic or Latino patients, as well as Hispanic or Latino Medicaid patients were less likely to have PCPs using EHRs, compared with privately-insured White patients, after controlling for PCPs’ practice type and location, as well as patient characteristics. This finding reflects a mixture of high and low EHR adopters among PCPs for poor minority patients.

Factors associated with ability to treat pediatric emergencies in US hospitals.

Objective: The purpose of this analysis is to investigate hospital and community factors associated with the availability of pediatric services, expertise, and supplies in US hospitals for treating pediatric emergencies. Methods: Data from the Emergency Pediatric Services and Equipment Supplement, a component of the 2002-2003 National Hospital Ambulatory Medical Care Survey, were merged with hospital and community characteristics to model preparedness to treat pediatric emergencies. The National Hospital Ambulatory Medical Care Survey samples nonfederal, short-stay, and general hospitals in the United States. The Emergency Pediatric Services and Equipment Supplement was based on the 2001 guidelines developed by the American Academy of Pediatrics and the American College of Emergency Physicians. Estimates were weighted to produce unbiased national estimates of pediatric services, expertise, and equipment availability in emergency departments. Logistic regression was used to model the probability of being better prepared based on the above guidelines. Results: Bivariate analyses showed that hospital inpatient pediatric structure was linearly related to availability of supplies. However, inpatient structure was not associated with presence of a pediatric trauma service or written transfer agreement. Logistic regressions with each preparedness measure indicated that, after adjusting for hospital and community factors, pediatric volume, teaching hospital status, geographic region, and per capita income of the community were strongly related to being better prepared on each of the preparedness measures. Conclusions: To meet the 2001 guidelines, emergency departments need to improve their inventory of pediatric supplies, and hospitals that do not have specialized inpatient services need to implement written transfer agreements with other hospitals.

Trends in diabetes treatment patterns among primary care providers.

Using data from the National Ambulatory Medical Care Survey, logit models tested for trends in the probability that visits by adult diabetes patients to their primary care providers included recommended treatment measures, such as a prescription for an angiotensin-converting enzyme (ACE) inhibitor or an angiotensin-II receptor blocker (ARB), blood pressure measurement, and diet/nutrition or exercise counseling. Results indicated that the probability that visits included prescription of an ACE or ARB and blood pressure measurement increased significantly over the 1997–2005 period, while the probability that visits documented provision of exercise counseling rose since 2001.