Survey
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
* Your assessment is very important for improving the workof artificial intelligence, which forms the content of this project
1991 NAMCS DATA FILE DOCUMENTATION DESCRIPTION OF THE NATIONAL AMBULATORY MEDICAL CARE SURVEY INTRODUCTION This data file comprises the data collected by the National Ambulatory Medical Care Survey (NAMCS) in 1991, conducted by the National Center for Health Statistics (NCHS). The National Ambulatory Medical Care Survey provides data from samples of patient records selected from a national sample of office-based physicians. These national estimates describe the utilization of ambulatory medical care services in the United States. In 1991 there were 33,795 patient records provided by 1,354 doctors that participated in the survey. For a brief description of the survey design and data collection procedures, see below. For a more detailed description of the survey design, data collection procedures, and the estimation process see references 1 and 2. There are two important points relative to analyzing data from this file that should be noted: 1-Data users should be fully aware of the importance of the "patient visit weight" and how it must be used. Information about the patient visit weight is presented on page 13. If more information is needed the staff of the Ambulatory Care Statistics Branch can be consulted by calling 301/436-7132 during regular working hours. 2-Users should also be aware of the reliability or unreliability of certain estimates, particularly the smaller estimates. The National Center for Health Statistics, which conducts many health surveys, considers an estimate to be reliable if it has a relative standard error of 30% or less. Therefore, it is important to know the value of the lowest possible estimate in this survey that is considered reliable, so as not to present data in a journal article or paper that may be unreliable. Most data file users can obtain an adequate working knowledge of the relative standard error from the information presented in the Appendix. If you would like more information, do not hesitate to consult the staff of the Ambulatory Care Statistics Branch. SCOPE OF THE SURVEY The basic sampling unit for the NAMCS is the physician-patient encounter or visit. Only visits in the offices of nonfederally employed physicians classified by the American Medical Association (AMA) or the American Osteopathic Association (AOA) as "office-based, patient care" were included in the 1991 NAMCS. In addition, physicians in the specialties of anesthesiology, pathology, and radiology were excluded from the physician universe. - 1 - Major types of ambulatory encounters not included in the 1991 NAMCS were those made by telephone, those made outside of the physician's office, and those made in hospital or institutional settings. SAMPLING FRAME AND SIZE OF SAMPLE The sampling frame for the 1991 NAMCS was composed of all physicians contained in the master files maintained by the AMA and AOA as of December 31, 1990, who met the following criteria: Office-based, as defined by the AMA and AOA: Principally engaged in patient care activities; Nonfederally employed; Not in specialties of anesthesiology, pathology, and radiology. The 1991 NAMCS sample included 2,540 physicians: 2,276 MD's and 264 Doctors of Osteopathy. Sample physicians were screened at the time of the survey to assure that they met the above-mentioned criteria, 653 physicians did not meet all of the criteria and were, therefore, ruled out of scope (ineligible) for the study. The most frequent reasons for being out of scope were that the physician was retired, deceased, or employed in teaching, research, or administration. Of the 1,887 inscope (eligible) physicians, 72 percent participated in the study. The physician universe, sample size, and response rates by physician specialty are shown in table I. Of the participating physicians, 216 saw no patients during their assigned reporting period because of vacations, illness, or other reasons for being temporarily not in practice. - 2 - TABLE I. DISTRIBUTION OF PHYSICIANS IN THE UNIVERSE AND IN THE 1991 NATIONAL AMBULATORY MEDICAL CARE SURVEY SAMPLE BY PHYSICIAN SPECIALTY AND RESPONSE STATUS: UNITED STATES, 1991. _________________________________________________________________________ _____ NAMCS OUT OF IN NON-RES- RES- RESPONSE UNIVERSE* SAMPLE SCOPE SCOPE PONSE PONSE 533 1,354 RATE** ALL SPECIALTIES 72 334,714 2,540 GENERAL/FAMILY PRACTICE 71 57,754 364 115 249 73 176 OSTEOPATHY 75 14,558 264 58 206 51 155 INTERNAL MEDICINE 63 46,891 206 56 150 56 94 PEDIATRICS 77 28,117 177 65 112 26 86 GENERAL SURGERY 78 19,356 229 62 167 36 131 OBSTETRICS & GYNECOLOGY 72 25,422 171 37 134 38 96 ORTHOPEDIC SURGERY 69 14,237 116 8 108 34 74 CARDIOVASCULAR DISEASES 60 11,027 125 28 97 39 58 DERMATOLOGY 80 6,217 101 16 85 17 68 UROLOGICAL SURGERY 71 7,366 117 18 99 29 70 23,058 105 38 67 14 53 5,758 98 15 83 28 55 13,149 108 23 85 22 63 PSYCHIATRY 79 NEUROLOGY 66 OPHTHALMOLOGY 74 653 1,887 OTOLARYNGOLOGY 68 ALL OTHER SPECIALTIES 73 6,433 105 20 85 27 58 55,371 254 94 160 43 117 _________________________________________________________________________ ______ *THESE DATA ARE DERIVED FROM THE AMERICAN MEDICAL ASSOCIATION AND THE AMERICAN OSTEOPATHIC ASSOCIATION AND REPRESENT THE TOTAL NUMBER OF PHYSICIANS WHO ARE ELIGIBLE FOR THE NAMCS. ** RESPONSE RATE (PERCENT) = RESPONSE DIVIDED BY IN SCOPE. - 3 - SAMPLE DESIGN The 1991 NAMCS utilized a multistage probability design that involved probability samples of primary sampling units (PSU's), physician practices within PSU's, and patient visits within practices. The first-stage sample included 112 PSU's. A PSU is a county, a group of adjacent counties, or a standard metropolitan statistical area (SMSA). The second stage consisted of a probability sample of practicing physicians selected from the master files maintained by the American Medical Association (AMA) and American Osteopathic Association (AOA). Within each PSU, all eligible physicians were stratified by fifteen specialty groups: general and family practice, osteopathy, internal medicine, pediatrics, general surgery, obstetrics and gynecology, orthopedic surgery, cardiovascular diseases, dermatology, urological surgery, psychiatry, neurology, ophthalmology, otolaryngology, and all other specialties. The final stage was the selection of patient visits within the annual practices of sample physicians. This involved two steps. First, the total physician sample was divided into 52 random subsamples of approximately equal size, and each subsample was randomly assigned to 1 of the 52 weeks in the survey year. Second, a systematic random sample of visits was selected by the physician during the assigned week. The sampling rate varied for this final step from a 100-percent sample for very small practices to a 20-percent sample for very large practices as determined in a presurvey interview. The method by which the sampling rate was determined is described in reference 3. FIELD ACTIVITIES The first contact with the sample physician is through a letter from the Director, NCHS. This may be accompanied by a letter from one of the 16 national medical associations that endorse the NAMCS, providing the physician is a member of one or more of these associations. After the physician has received the introductory letter(s) the field representative telephones the physician to set up an appointment with him or her to discuss the survey and instruct the doctor on how to complete the forms. DATA COLLECTION The actual data collection for the NAMCS was carried out by the physician aided by his or her office staff when possible. Two data collection forms were employed by the physician: The Patient Log and the Patient Record. The Patient Log is used to sequentially list patients seen in the physician's office during his assigned reporting week. This list served as the sampling frame to indicate the visit for which data were to be recorded. A perforation between the patient names and patient visit characteristics permitted the physician to remove patient names and protect confidentiality. - 4 - Based on the physician's estimate of the expected number of office visits, each physician was assigned a patient-sampling ratio. These ratios were designed so that about 30 Patient Records were completed during the assigned reporting week. Physicians expecting 10 or fewer visits each day recorded data for all of them, while those expecting more than 10 visits per day recorded data for every second, third, or fifth visit based on the predetermined sampling interval. These procedures minimized the data collection workload and maintained approximately equal reporting levels among sample physicians regardless of practice size. For physicians assigned a patient sampling ratio, a random start was provided on the first page of the log, so that predesignated sample visits on each succeeding page of the log provided a systematic random sample of patient visits during the reporting period. DATA PROCESSING In addition to the completeness checks made by the field staff, clerical edits were performed upon receipt of the data for central processing. These procedures proved quite efficient, reducing the item nonresponse rates to a negligible amount--5 percent or less for most data items. For items 5, 9, 13, 15, 16, and 19 on the Patient Record form, nonresponse rates ranged from 6 to 8 percent. Information from the Induction Interview and Patient Record forms was keypunched with 100-percent verification and converted to electronic data file. Extensive computer consistency and edit checks were also performed. Data items still unanswered at this point were imputed by randomly assigning a value from a Patient Record with similar characteristics; imputations were based on physician specialty and broad diagnostic categories. MEDICAL CODING The Patient Record form contains four medical items requiring three separate coding systems. The three coding systems are described briefly below. A two-way independent verification procedure was used to control the medical coding operation. Differences between coders were adjudicated by a third coder. (A) Patient's Reason for Visit: Information contained in item 10 (patient's reason for visit) of the Patient Record was coded according to an updated version of A Reason for Visit Classification for Ambulatory Care (RVC) (reference 4). The updated classification has not been published as of this writing but the list of codes is contained in Appendix II. The classification was updated to incorporate several new codes as well as changes to existing codes. The system continues to utilize a modular structure composed of seven reason-for-visit modules. The digits 1 through 7 precede the RVC codes to identify the various modules, and digits 8 and 9 are used to identify uncodable entries and blank entries as follows: - 5 - Prefix Module "1" = symptom module "2" = disease module "3" = diagnostic, screening, and preventive module "4" = treatment module "5" = injuries and adverse effects module "6" = test results module "7" = administrative module "8" = Uncodable entries "9" = special code = blank A maximum of 3 reasons for visit were coded in sequence; coding instructions concerning the patient's reason for visit are contained in the NAMCS Medical Coding Manual (reference 6). Copies are available upon request. The SETS version of the NAMCS data allows the user to select reason for visit codes at either the 3 digit level or the 4 digit level. The variable name VISIT3DG should be used to access codes at the 3 digit level; if greater specificity is needed, the variable name VISITRE will access codes at the 4 digit level. (B) Physician's Diagnoses: Diagnostic information in item 11 of the Patient Record was coded according to the International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM) (reference 5). For 1991, NAMCS micro-data file tabulations involving physician's diagnoses, the following characteristics exist: 1. The prefix "1" preceding the diagnostic codes has been added to identify all codes in the range 001.00-999.90, e.g. '138100' = '381.00' = Acute nonsupporative otitis media, unspecified. 2. The prefix "20" preceding diagnostic codes represents V code diagnoses V01.00-V82.90, e.g. '201081' = 'V10.81' = personal history of malignant neoplasm of bone. change In other words, the "20" to "V." NOTE: The use of prefixes facilitates the calculation of percent distributions, while substituting "20" for the letter "V" allows that all diagnostic fields on the data file will contain numerical data. - 6 - 3. In addition to the diagnostic codes from the ICD-9-CM there are 3 unique codes in the diagnostic fields that were developed by the NAMCS staff: 209900 = - noncodable diagnosis - insufficient information for coding - illegible diagnosis 209970 = diagnosis of "none" 900000 = blank diagnosis A maximum of three diagnoses were coded in sequence: coding instructions concerning diagnoses are contained in the NAMCS Medical Coding Manual (reference 6). The SETS version of the NAMCS data allows the user to select ICD-9-CM codes at the 3-digit level. The variable name DIAG3DGT should be used to access codes at the 3 digit level;if greater specificity is needed, use the variable name DIAGNOS to access codes at the highest level. (C) Ambulatory Surgical Procedures: Information on ambulatory diagnostic or therapeutic surgical procedures scheduled or performed at the current office visit is collected in item 14 of the Patient Record form. These data were also coded according to the International Classification of Diseases (ICD-9-CM), using the 4-digit surgical codes in Volume III. A maximum of two surgical procedures were coded for each visit. The SETS version of the NAMCS data allows the user to select ICD-9-CM codes for ambulatory surgical procedures at the 2-, 3, and 4-digit level. The variable name AMBSRG1 should be used to access codes for the first-listed ambulatory surgical procedure at the 4-digit level; use the variable SRG12DGT to examine codes at the 2-digit level and SRG13DGT for codes at the 3-digit level. Likewise, to access codes for the second-listed ambulatory surgical procedure, use variables names AMBSRG2, SRG22DGT, and SRG23DGT. (D) Medication Therapy This Visit: The NAMCS drug data in item 17 have been classified and coded according to a unique classification scheme (reference 7) developed at NCHS. A list of the drug codes are in Appendix III. The Patient Record form allows for the recording of up to 5 drugs. The file format includes a corresponding allocation of drug fields. - 7 - A report describing the method and instruments used to collect and process drug information for the NAMCS has been published (reference 8). Copies are available upon request. It should be noted that a substantial amount of information concerning each drug is not contained in these data. This additional drug information will be available in CD-ROM format beginning with the 1992 NAMCS. For years prior to the 1992 NAMCS, the additional drug information is available as a public-use data tape, or special tabulations may be requested from the Ambulatory Care Statistics Branch. For more information on the availability of the NAMCS drug data, please contact the Ambulatory Care Statistics Branch, Division of Health Care Statistics, Room 9-52, 6525 Belcrest Road, Hyattsville, Maryland 20782. The telephone number is 301/436-7132. HOW TO SELECT PRINCIPAL REASONS, DIAGNOSES, AND MEDICATIONS The SETS version of the NAMCS allows the user to examine all-listed occurrences within a single visit for the following variables: Reason for visit; 3-digit reason; Diagnosis; 3-digit diagnosis; Medications; by the following variable names: VISITRE; VISIT3DG; DIAGNOS; DIAG3DGT; MEDCODE. If the user is interested in principal entries within any of these variables, then a one in brackets [1] should be added to the variable name. Hence, the principal diagnosis would be DIAGNOS[1]. The same logic is true for specifying the 2nd or 3rd diagnosis or 4th or 5th medication listed. Similarly, when the user is interested in further information about the medication codes, then NEWMED with a number in brackets will provide information on whether the medication is new or continuing (see page 9). NOTE: The user should be aware of the following conditions when analyzing records for which there may be multiple occurrences of a particular code. For example, this situation might occur when the user is looking at all occurrences of a particular 3-digit diagnosis. Because up to 3 diagnoses can be coded per record, it is possible that a single record may contain multiple occurrences of a diagnosis at the 3-digit level. An example is a patient who has both an acute (382.00) and a chronic (382.1) form of otitis media recorded at the office visit. When performing a simple query in SETS using unweighted NAMCS data, the result will reflect all occurrences of the condition of interest. This means that multiple occurrences of a 3-digit diagnostic code appearing on a single record will each be counted in the total that is output by the SETS program. However, if the user applies patient weights to the query variable or otherwise links it to a single-field variable on the record, SETS will output a count that reflects all RECORDS containing the condition of interest. This means that a record that contains more than one occurrence of a 3-digit diagnostic code will only be counted once. The same process occurs when the user creates tables in SETS; only records are counted, not multiple occurrences of a condition on each record. - 8 - As a result, if the user performs a query on, say, DIAG3DGT = 1382, using weighted NAMCS data, the result may be smaller than what would be obtained by summing the results of separate queries for DIAG3DGT[1], DIAG3DGT[2], and DIAG3DGT[3]. This would apply in cases where there are multiple occurrences of that 3-digit diagnosis on a single record. An additional caveat to users concerns the NEWMED variable. In order to obtain correct results using this variable, it is necessary to link it both with a medication code of interest AND with the appropriate position (whether it is the 1st, 2nd, 3rd, 4th, or 5th medication listed) on the record. For example, in order to determine how many times a particular drug was mentioned as a new medication or as a continued medication, it is necessary to analyze data for each of the five MEDCODES on the record individually and then to sum the results. Selecting a MEDCODE of interest and creating a table for NEWMED without linking the position of each on the record will produce incorrect results. In order to do this, the user would first determine the code for the medication of interest (found in Appendix III), create a subset for MEDCODE[1] = the code of interest. Then create a table using the variable name NEWMED[1] to determine how often it was a new or continued medication. This process would be repeated for the 2nd, 3rd, 4th, and 5th positions of MEDCODE on the record, each time specifying the corresponding position of NEWMED. The user could then sum the results to determine how many times the drug was prescribed as a new or continued medication. POPULATION FIGURES The base population used in computing annual visit rates is presented in table II. These figures are based on provisional estimates for the civilian noninstitutionalized population as of July 1, 1991, provided by the U.S. Bureau of the Census. The population estimates are presented here solely for the purpose of providing denominators for rate computation and should not be considered as official population estimates. IMPORTANT: The SETS program allows the user to obtain visit rates that are age-specific, sex-specific, and race-specific. When requesting visit rates by race, however, the recoded race variable RACER must be used rather than the original race variable. If visit rates are requested for NAMCS variables other than age, sex, and race, the SETS program will calculate them using the total U.S. population as the denominator. For this reason, users who wish to obtain rates by geographic region, using regional population estimates as denominators, will need to refer to table II and calculate them manually. The SETS program produces visit rates by age for single years unless otherwise specifed. To define age groups, use the following syntax at the edit prompt in the table option with the age groups of interest: EXAMPLE: age (>15, 15 to 24, 25 to 44, 45 to 64, >64) - 9 - Table II. rates U.S. population 1/ estimates used in computing annual visit for the National Ambulatory Care Survey, by age, sex, race, and geographic region: July 1, 1991. ___________________________________________________________________ | | Population Age |_________________________________________________ Race, Sex, | | | | | and Region | All | < 5 | 5-14 | 15-24 | 25-34 | Ages | years | years | years | years ___________________________________________________________________ (in 1,000's 2/) Total | 248,712 | 19,379 | 36,249 | 34,534 | 42,404 | 120,724 | | 127,988 | 9,918 | 9,461 | 18,557 | 17,692 | 17,131 | 17,403 | 20,871 21,533 | 208,202 | | 101,586 | | 106,616 | 15,525 | 7,931 | 7,594 | 29,182 | 14,953 | 14,229 | 27,726 | 13,845 | 13,881 | 35,091 17,477 17,614 Sex Male Female Race and Sex White Male Female Black Male Female | | | 30,896 | 14,435 | 16,461 | 3,085 | 1,569 | 1,516 | 5,671 | 2,883 | 2,788 | 5,077 | 2,428 | 2,649 | 5,357 2,430 2,927 Other Male Female | | | 9,614 | 4,704 | 4,911 | 770 | 419 | 351 | 1,396 | 721 | 675 | 1,731 | 858 | 873 | 1,957 965 992 Region Northeast Midwest South West | | | | 50,300 59,735 84,008 54,670 - 10 - Table II. rates U.S. population 1/ estimates used in computing annual visit for the National Ambulatory Care Survey, by age, sex, race, and geographic region: July 1, 1991--Con. ____________________________________________________________________ | | Population Age Race, Sex, |__________________________________________________ and Region | 35-44 | 45-54 | 55-64 | 65-74 | 75 years | years | years | years | years | and over ____________________________________________________________________ (in 1,000's 2/) Total | 38,693 | 26,004 | 21,157 | 18,301 | 11,991 | | 18,970 | 19,723 | 12,611 | 13,393 | 10,015 | 11,142 | 8,183 | 10,118 | 4,469 7,522 Race and Sex White Male Female | | | 32,854 | 16,333 | 16,521 | 22,261 | 10,913 | 11,348 | 18,367 | 8,729 | 9,638 | 16,323 | 7,348 | 8,975 | 10,874 4,058 6,816 Black Male Female | | | 4,283 | 1,916 | 2,367 | 2,707 | 1,198 | 1,509 | 2,155 | 974 | 1,181 | 1,609 | 694 | 915 | 954 344 610 Other Male Female | | | 1,556 | 721 | 835 | 1,037 | 500 | 537 | 635 | 312 | 323 | 368 | 141 | 227 | 163 67 96 Sex Male Female Region Northeast | Midwest | South | West | ____________________________________________________________________ 1/ Civilian noninstitutional population 2/ Numbers may not add to totals because of rounding. - 11 - ESTIMATION PROCEDURES Statistics produced from the 1991 National Ambulatory Medical Care Survey were derived by a multistage estimating procedure. The procedure produces essentially unbiased national estimates and has basically three components: (1) inflation by reciprocals of the probabilities of selection, (2) adjustment for nonresponse, and (3) a ratio adjustment to fixed totals. Each of these components is described briefly below. (1) INFLATION BY RECIPROCALS OF SAMPLING PROBABILITIES.-- Since the survey utilized a three-stage sample design, there were three probabilities: (A) The probability of selecting the PSU, (B) the probability of selecting a physician within the PSU, and (C) the probability of selecting a patient visit within the physician's practice. The last probability was defined to be the exact number of office visits during the physician's specified reporting week divided by the number of Patient Records completed. All weekly estimates were inflated by a factor of 52 to derive annual estimates. (2) ADJUSTMENT FOR NONRESPONSE--Estimates from the NAMCS data were adjusted to account for sample physicians who did not participate in the study. This was done in such a manner as to minimize the impact of nonresponse on final estimates by imputing to nonresponding physicians the practice characteristics of similar responding physicians. For this purpose, similar physicians were judged to be physicians having the same specialty designation and practicing in the same PSU. (3) RATIO-ADJUSTMENT.--A poststratification adjustment was made within each of fifteen physician specialty groups. The ratio adjustment is a multiplication factor which had as its numerator the number of physicians in the universe in each physician specialty group, and as its denominator the estimated number of physicians in that particular specialty group. The numerator was based on figures obtained from the AMA-AOA master files, and the denominator was based on data from the sample. SAMPLING ERRORS Procedures for calculating sampling errors as well as estimates of standard errors of statistics derived from the NAMCS are described in the technical notes of reference 2, as well as Appendix I of this document. - 12 - PATIENT VISIT WEIGHT The "patient visit weight" is a vital component in the process of producing national estimates from sample data and its use should be clearly understood by all data file users. The statistics contained on the data file reflect data concerning only a sample of patient visits--and not a complete count of all the visits that occurred in the United States. Each record on the data file represents one visit in the sample of 33,795 visits. In order to obtain national estimates from the sample, each record is assigned an inflation factor called the "patient visit weight." By aggregating the "patient visit weights" on the 33,795 sample records for 1991 the user can obtain the total of 669,689,374 estimated visits made by all patients in the United States. Marginal tables which may be selected from the main menu present numbers of records for selected variables as well as the corresponding national estimates of visits obtained by aggregating the "patient visit weights" on those records. PHYSICIAN-PATIENT LINKING CODE The purpose of this code is to allow for greater analytical depth by permitting the user to link individual Patient Record forms on the micro-data file with individual physicians' practices. This linking code will enable users to conduct more comprehensive analyses without violating the confidentiality of patients or physicians. The patient visit weight should never be applied directly to the physicianpatient linking code. - 13 - REFERENCES 1/ 1. National Center for Health Statistics: National Ambulatory Medical Care Survey: background and methodology, United States. National Center for Health Statistics. Vital Health Stat. 2(61). 1974. 2. Schappert SM. National Ambulatory Medical Care Survey, 1991 Summary. National Center for Health Statistics. Vital Health Stat. 13(116.) 1994. 3. Induction Interview Form. National Ambulatory Medical Care Survey. Bureau of the Census. OMB NO. 0920-0234. 4. Schneider D, Appleton L, McLemore T. A Reason for Visit Classification for Ambulatory Care. National Center for Health Statistics. Vital Health Stat. 2(78). 1979. 5. Public Health Service and Health Care Financing Administration. International Classification of Diseases, 9th Revision, Clinical Modification (ICD-9-CM). Washington: Public Health Service. 1980. 6. National Ambulatory Medical Care Survey: Medical Coding Manual. Updated annually. 7. National Ambulatory Medical Care Survey: MEDICATION CODE LIST. Updated annually. 8. Koch H, Campbell W. The collection and processing of drug information. National Ambulatory Medical Care Survey, 1980. National Center for Health Statistics. Vital Health Stat. 2(90). 1982. l/Information concerning additional reports using NAMCS data that have been published or are scheduled for publication through NCHS may be obtained from the Ambulatory Care Statistics Branch. - 13 -