TY - JOUR
AU - Groseclose, Samuel L.
AB - Abstract Despite state and local laws requiring medical providers to report notifiable infectious diseases to public health authorities, reporting is believed to be incomplete. Through means of an analytical literature review, the authors synthesize current knowledge on the completeness of disease reporting and identify factors associated with reporting completeness. The review was limited to published studies, conducted in the United States between 1970 and 1999, that quantitatively assessed infectious disease reporting completeness. Thirty-three studies met the inclusion criteria. Reporting completeness, expressed between 0% and 100%, was treated as the dependent outcome variable in statistical analysis; disease, study location, time period, study design, and study size were treated as independent variables. Fifty-six distinct measures of reporting completeness were identified for 21 diseases. Reporting completeness varied from 9% to 99% and was most strongly associated with the disease being reported. The mean reporting completeness for acquired immunodeficiency syndrome, sexually transmitted diseases, and tuberculosis as a group was significantly higher (79%) than for all other diseases combined (49%) (p < 0.01). communicable disease control, disease notification, population surveillance, review literature AIDS, acquired immunodeficiency syndrome Surveillance for infectious diseases is a critical element in providing effective public health disease control and prevention services. In the United States, disease reporting is mandated by state and local laws. These laws require medical providers and laboratories to notify state or local public health authorities of patients diagnosed with reportable conditions (1). A reportable condition is one for which “… regular, frequent, and timely information regarding indi-vidual cases is considered necessary for the prevention and control of the disease” (2, p. v). On the basis of the unique disease control priorities within each state, the state determines which conditions are reportable within its jurisdiction. States then voluntarily report diseases notifiable at the national level to the federal Centers for Disease Control and Prevention, and a summary of notifiable disease activity in the United States is published annually (2). Although this national system of disease surveillance probably results in incomplete case enumeration, it is less clear what proportion of notifiable disease events are captured by this confluence of local, state, and national public health surveillance activities. Periodic evaluation of disease surveillance activities may include quantitative measurements of the sensitivity of systems to detect conditions under surveillance (3). The sensitivity of surveillance systems includes both a case detection and diagnosis component and a disease-reporting component. Disease-reporting completeness refers to the proportion of those diagnosed with a notifiable condition that are reported to the appropriate public health authorities. Disease reporting, if it is representative and consistent over time, need not be complete to successfully monitor demographic, spatial, and temporal trends (4). Completeness becomes more important for infrequently occurring diseases, however, and some measure of reporting completeness is necessary to accurately interpret disease incidence or to make national and international comparisons among public health jurisdictions (5). Several studies have evaluated the completeness of reporting for a particular disease, in a particular jurisdiction, and over a particular period. We review here the published literature on the completeness of notifiable infectious disease reporting in the United States. Through means of an analytical literature review, we attempt to synthesize current knowledge on the completeness of disease reporting and to identify factors associated with reporting completeness. MATERIALS AND METHODS We reviewed all studies of notifiable infectious disease-reporting completeness that were published between 1970 and 1999 in the peer-reviewed biomedical and public health literature. Meeting abstracts and non-peer-reviewed publications were not included. Reports were identified by searching the MEDLINE database (US National Library of Medicine, from 1966 to April 2000) using the Internet Grateful Med (version 2.6.3) search engine. Several query combinations were used that involved the subject keywords “surveillance,” “evaluation,” “reporting,” and “completeness.” Additional reports were identified by manually reviewing the reference lists of those studies previously retrieved. The review was limited to studies conducted in the United States and its overseas territories that provided quantitative measurements of the completeness of notifiable infectious disease reporting to local, state, or national public health authorities. Evaluations included in this review must have assessed reporting completeness by comparing the number of notifiable disease reports received through conventional disease-reporting mechanisms with the total number of such disease events identified or estimated through the use of one or more additional data sources. Furthermore, the conventional and supplemental data sources must have covered the same geographic area and time period, thus controlling for true spatial and temporal fluctuations in disease incidence. Studies relying exclusively on physician-recall methods were not included in our review. Such studies have attempted to measure reporting completeness by asking providers to estimate the number of cases of a particular disease they have treated during a given time period and have compared this value with the number of cases reported (6, 7). Such studies suffer methodological limitations, however, because enumeration of incident cases based solely on physician recall has been shown to overestimate the true number of cases (8, 9). Information abstracted from each study included the disease or diseases under surveillance, the geographic area and time covered, the data sources and database matching criteria, the data validation methods used, the study design and statistical methods used, and the measured completeness of reporting. Some studies included more than one measurement of disease-reporting completeness, based on comparison with multiple supplemental data sources. In such instances, each distinct measure of reporting completeness was recorded from the study. Other important indicators of the quality of surveillance data (e.g., timeliness, representativeness) were not uniformly addressed by all studies and, therefore, are beyond the scope of this review. Reporting completeness, expressed as a continuous variable between 0 percent and 100 percent, was treated as the dependent outcome variable in multivariable analysis with disease, study location, time period, study design, and study size treated as independent variables. Study location was expressed by census regions defined by the US Census Bureau (Northeast, Midwest, South, West). Additional categories were created for Puerto Rico and for studies conducted nationally. Time was expressed as an ordinal variable by the decade (1970s, 1980s, or 1990s) in which the majority of data for each study were collected. Study design was divided into two categories to account for differences between uncorrected and underascertainment corrected methods for measuring reporting completeness (10, 11). Study size was expressed as a continuous variable by the denominator of the fraction used to calculate reporting completeness. Only results reported for discrete disease entities (rather than disease categories such as viral hepatitis), with a minimum of 10 cases reported in the denominator, were included in the statistical analysis. Statistical analyses were performed using SAS JMP version 3.1 software for Windows (SAS Institute, Inc., Cary, North Carolina). Finally, we divided the measures of reporting completeness observed in this review into two broad disease categories for comparison. The first category included the observations of reporting completeness for tuberculosis, acquired immunodeficiency syndrome (AIDS), and sexually transmitted diseases (i.e., chlamydia, gonorrhea, and syphilis). We grouped these conditions because surveillance and prevention activities for tuberculosis, AIDS, and sexually transmitted diseases are often grouped programmatically, since the populations affected by these conditions and the surveillance methods and data sources are similar. The second category included observations of reporting completeness for all other disease entities for which data were available. RESULTS A total of 33 studies were identified in the published literature that met the stated inclusion criteria. Among the published studies, we observed two distinct methods for calculating reporting completeness. In 22 studies (67 percent), researchers measured reporting completeness by dividing the number of cases reported to public health authorities by the total number of cases identified through active case detection and the use of supplemental data sources (table 1) (8, 12–32). This method does not account for the number of cases left undetected by all data sources and is referred to as the uncorrected method. In the remaining 11 studies, researchers measured reporting completeness by dividing the number of cases reported to public health authorities by the total number of cases estimated through the use of capture-recapture methods for comparing two or more data sources (table 2) (33–43). This method attempts to account for the number of cases undetected by all available data sources and is referred to as the underascertainment corrected method (44). When determining the accuracy of this method, one should consider the validity of the data from each source, the dependency relation between data sources, and the criteria used to match persons between data sources. Specific observations related to these factors are noted for each study in the limitations column of table 2. TABLE 1. Published reports measuring completeness of notifiable disease reporting: uncorrected methods Author(s) (reference no.)  Location  Time period  Disease  Supplemental data sources  Observed reporting completeness*  Marier (12)  Washington, DC  1971  Various  Hospital discharge records  Total†: 35% (200 of 570)            Haemophilus influenzae: 32% (7 of 22)            Salmonellosis: 42% (11 of 26)            Shigellosis: 62% (21 of 34)            Tuberculosis: 63% (127 of 200)  Murray et al. (13)  Nassau County, NY  1972  Tuberculosis  Laboratory data  65% (32 of 49)  Eisenberg and Wiesner (8)  Alaska  1973–1974  Gonorrhea  Medical records  42% (76 of 183)  Kimball et al. (14)  Washington, DC  1978  Shigellosis  Medical records  30% (43 of 136)  Davis and Bohn (15)  Wisconsin  1980–1982  Meningococcal disease  Laboratory data  23% (28 of 120)  Vogt et al. (16)  Vermont  1982–1983  Various  Hospital discharge records  Total†: 40% (144 of 364)            Hepatitis A and B: 44% (17 of 39)            Gonorrhea: 93% (28 of 30)            Pertussis: 40% (6 of 15)            Salmonellosis: 67% (42 of 63)  Chamberland et al. (17)  New York, NY  1982–1983  AIDS‡  Medical records and laboratory data combined  96% (409 of 425)  Harkess et al. (18)  Oklahoma  1985  Shigellosis  Laboratory data  81% (99 of 122)  Lindan et al. (19)  San Francisco, CA  1985–1986  AIDS  Death certificates  92% (1,171 of 1,273)  Conway et al. (20)  South Carolina  1986–1987  AIDS  Hospital discharge records  60% (91 of 153)  Modesitt et al. (21)  Oregon  1986–1987  AIDS  Hospital discharge records and death certificates combined  66% (56 of 85)  Campos-Outcalt et al. (22)  Pima County, AZ  1986–1988  Various  Provider group medical billing records  Total†: 64% (183 of 286)            Hepatitis B: 50% (10 of 20)            Pertussis: 64% (9 of 14)            Giardiasis: 10% (1 of 10)            Hepatitis A; 33% (4 of 12)            Shigellosis: 67% (10 of 15)            AIDS: 80% (92 of 115)            Gonorrhea: 95% (21 of 22)            Syphilis: 79% (23 of 29)            Tuberculosis: 40% (6 of 15)  Fife et al. (23)  Philadelphia, PA  1986–1991  AIDS  Provider-based case registry  91% (267 of 295)  Kirsch et al. (24)  Washington, DC  1989  Gonorrhea  Emergency department medical records and laboratory data combined  91% (204 of 223)  Elcock et al. (25)  San Mateo County, CA  1989–1991  AIDS  Medical records  81% (95 of 118)  Standaert et al. (26)  Tennessee  1989–1992  Invasive bacterial infections  Laboratory data  Neisseria meningitidis: 58% (41 of 71)            H. influenzae: 49% (94 of 191)  Ewert et al. (27)  Los Angeles, CA  1990–1991  Measles  Community survey  63% (10 of 16)  Driver et al. (28)  Puerto Rico  1992  Tuberculosis  Hospital discharge records, pharmacy records, laboratory data, and case registries combined  86% (136 of 159)  Yokoe et al. (29)  Massachusetts  1992–1996  Tuberculosis  Automated clinical, laboratory, and pharmacy data from health maintenance organization  81% (35 of 43)  Smucker and Thomas (30)  Rural North Carolina  1993  Sexually transmitted diseases  Laboratory data compared with case reports during both active and passive surveillance periods  Gonorrhea             72% during passive surveillance (80 of 111)             88% during active surveillance (21 of 24)            Chlamydia             55% during passive surveillance (87 of 158)             79% during active surveillance (19 of 24)  Dembek et al. (31)  Connecticut  1994–1996  Vancomycin-resistant enterococci  Laboratory data  59% (158 of 266)  Trepka et al. (32)  Wisconsin  1995  Tuberculosis  Hospital discharge records and laboratory data  98% (50 of 51 from discharge data)            99% (87 of 88 from laboratory data)  Author(s) (reference no.)  Location  Time period  Disease  Supplemental data sources  Observed reporting completeness*  Marier (12)  Washington, DC  1971  Various  Hospital discharge records  Total†: 35% (200 of 570)            Haemophilus influenzae: 32% (7 of 22)            Salmonellosis: 42% (11 of 26)            Shigellosis: 62% (21 of 34)            Tuberculosis: 63% (127 of 200)  Murray et al. (13)  Nassau County, NY  1972  Tuberculosis  Laboratory data  65% (32 of 49)  Eisenberg and Wiesner (8)  Alaska  1973–1974  Gonorrhea  Medical records  42% (76 of 183)  Kimball et al. (14)  Washington, DC  1978  Shigellosis  Medical records  30% (43 of 136)  Davis and Bohn (15)  Wisconsin  1980–1982  Meningococcal disease  Laboratory data  23% (28 of 120)  Vogt et al. (16)  Vermont  1982–1983  Various  Hospital discharge records  Total†: 40% (144 of 364)            Hepatitis A and B: 44% (17 of 39)            Gonorrhea: 93% (28 of 30)            Pertussis: 40% (6 of 15)            Salmonellosis: 67% (42 of 63)  Chamberland et al. (17)  New York, NY  1982–1983  AIDS‡  Medical records and laboratory data combined  96% (409 of 425)  Harkess et al. (18)  Oklahoma  1985  Shigellosis  Laboratory data  81% (99 of 122)  Lindan et al. (19)  San Francisco, CA  1985–1986  AIDS  Death certificates  92% (1,171 of 1,273)  Conway et al. (20)  South Carolina  1986–1987  AIDS  Hospital discharge records  60% (91 of 153)  Modesitt et al. (21)  Oregon  1986–1987  AIDS  Hospital discharge records and death certificates combined  66% (56 of 85)  Campos-Outcalt et al. (22)  Pima County, AZ  1986–1988  Various  Provider group medical billing records  Total†: 64% (183 of 286)            Hepatitis B: 50% (10 of 20)            Pertussis: 64% (9 of 14)            Giardiasis: 10% (1 of 10)            Hepatitis A; 33% (4 of 12)            Shigellosis: 67% (10 of 15)            AIDS: 80% (92 of 115)            Gonorrhea: 95% (21 of 22)            Syphilis: 79% (23 of 29)            Tuberculosis: 40% (6 of 15)  Fife et al. (23)  Philadelphia, PA  1986–1991  AIDS  Provider-based case registry  91% (267 of 295)  Kirsch et al. (24)  Washington, DC  1989  Gonorrhea  Emergency department medical records and laboratory data combined  91% (204 of 223)  Elcock et al. (25)  San Mateo County, CA  1989–1991  AIDS  Medical records  81% (95 of 118)  Standaert et al. (26)  Tennessee  1989–1992  Invasive bacterial infections  Laboratory data  Neisseria meningitidis: 58% (41 of 71)            H. influenzae: 49% (94 of 191)  Ewert et al. (27)  Los Angeles, CA  1990–1991  Measles  Community survey  63% (10 of 16)  Driver et al. (28)  Puerto Rico  1992  Tuberculosis  Hospital discharge records, pharmacy records, laboratory data, and case registries combined  86% (136 of 159)  Yokoe et al. (29)  Massachusetts  1992–1996  Tuberculosis  Automated clinical, laboratory, and pharmacy data from health maintenance organization  81% (35 of 43)  Smucker and Thomas (30)  Rural North Carolina  1993  Sexually transmitted diseases  Laboratory data compared with case reports during both active and passive surveillance periods  Gonorrhea             72% during passive surveillance (80 of 111)             88% during active surveillance (21 of 24)            Chlamydia             55% during passive surveillance (87 of 158)             79% during active surveillance (19 of 24)  Dembek et al. (31)  Connecticut  1994–1996  Vancomycin-resistant enterococci  Laboratory data  59% (158 of 266)  Trepka et al. (32)  Wisconsin  1995  Tuberculosis  Hospital discharge records and laboratory data  98% (50 of 51 from discharge data)            99% (87 of 88 from laboratory data)  * Observed reporting completeness is the number of routine surveillance case reports divided by the observed number of cases in routine and supplemental data sources combined. † Includes other notifiable conditions. ‡ AIDS, acquired immunodeficiency syndrome. View Large TABLE 2. Published reports measuring completeness of notifiable disease reporting: underascertainment-corrected methods Author(s) (reference no.)  Location  Time period  Disease  Supplemental data sources  Limitations  Estimated reporting completeness*  Cochi et al. (33)  United States  1970–1985  Congenital rubella syndrome  Birth defects registry  Personal identifiers not used for matching  22% (263 of 1,186)  Sutter et al. (34)  United States  1979–1984  Tetanus mortality  Death certificates  Outcome (alive/dead) variable in case reports not validated; personal identifiers not used for matching  40% (129 of 326)  Prevots et al. (35)  United States  1980–1991  Vaccine-associated paralytic polio  Claims filed with national vaccine injury compensation program  Matching criteria unspecified  81% (92 of 114)  Hardy et al. (36)  Boston, MA; Chicago, IL; Washington, DC; New York, NY  1985  AIDS†  Death certificates  Inclusion and search criteria varied across study sites  89% (487 of 548)  Sutter and Cochi (37)  United States  1985–1988  Pertussis  Hospital discharge records and death certificates  Hospitalization and outcome data from case reports not validated; discharge diagnoses and death certificates not validated; personal identifiers not used for matching  32% (4,404 of 13,557 hospitalized)              33% (32 of 98 deaths)  Rosenblum et al. (38)  Various sites, United States  1988  AIDS  Hospital discharge and Medicaid data combined  Matching criteria unspecified  92% (4,157 of 4,500)  Davis et al. (39)  New York, NY  1991  Measles  Medical records  No validation of physician diagnosis  45% (664 of 1,487)  Ikeda et al. (40)  New York State, excluding New York City  1991  Perinatal hepatits B exposure  Mother/infant pairs matched from maternal prenatal, maternal perinatal, and newborn infant screening program data  Matching criteria unspecified  96% (363 of 378)  Ackman et al. (41)  New York State, excluding New York City  1991  Meningococcal disease  Hospital discharge records    93% (100 of 107)  Barat et al. (42)  Southwestern United States  1995  Malaria  Laboratory data  Matching criteria unspecified  69% (43 of 62)  Effler et al. (43)  Hawaii  1998  Various  Automated laboratory data  Matching criteria unspecified  Total‡: 44% (156 of 357)              Salmonellosis: 50% (102 of 205)              Shigellosis: 54% (16 of 30)              Giardiasis: 33% (26 of 79)              VRE†: 22% (7 of 32)              Invasive Streptococcus pneumoniae: 9% (5 of 55)  Author(s) (reference no.)  Location  Time period  Disease  Supplemental data sources  Limitations  Estimated reporting completeness*  Cochi et al. (33)  United States  1970–1985  Congenital rubella syndrome  Birth defects registry  Personal identifiers not used for matching  22% (263 of 1,186)  Sutter et al. (34)  United States  1979–1984  Tetanus mortality  Death certificates  Outcome (alive/dead) variable in case reports not validated; personal identifiers not used for matching  40% (129 of 326)  Prevots et al. (35)  United States  1980–1991  Vaccine-associated paralytic polio  Claims filed with national vaccine injury compensation program  Matching criteria unspecified  81% (92 of 114)  Hardy et al. (36)  Boston, MA; Chicago, IL; Washington, DC; New York, NY  1985  AIDS†  Death certificates  Inclusion and search criteria varied across study sites  89% (487 of 548)  Sutter and Cochi (37)  United States  1985–1988  Pertussis  Hospital discharge records and death certificates  Hospitalization and outcome data from case reports not validated; discharge diagnoses and death certificates not validated; personal identifiers not used for matching  32% (4,404 of 13,557 hospitalized)              33% (32 of 98 deaths)  Rosenblum et al. (38)  Various sites, United States  1988  AIDS  Hospital discharge and Medicaid data combined  Matching criteria unspecified  92% (4,157 of 4,500)  Davis et al. (39)  New York, NY  1991  Measles  Medical records  No validation of physician diagnosis  45% (664 of 1,487)  Ikeda et al. (40)  New York State, excluding New York City  1991  Perinatal hepatits B exposure  Mother/infant pairs matched from maternal prenatal, maternal perinatal, and newborn infant screening program data  Matching criteria unspecified  96% (363 of 378)  Ackman et al. (41)  New York State, excluding New York City  1991  Meningococcal disease  Hospital discharge records    93% (100 of 107)  Barat et al. (42)  Southwestern United States  1995  Malaria  Laboratory data  Matching criteria unspecified  69% (43 of 62)  Effler et al. (43)  Hawaii  1998  Various  Automated laboratory data  Matching criteria unspecified  Total‡: 44% (156 of 357)              Salmonellosis: 50% (102 of 205)              Shigellosis: 54% (16 of 30)              Giardiasis: 33% (26 of 79)              VRE†: 22% (7 of 32)              Invasive Streptococcus pneumoniae: 9% (5 of 55)  * Estimated reporting completeness is the number of routine surveillance case reports divided by the estimated number of cases, after correcting for underascertainment using capture-recapture methods. † AIDS, acquired immunodeficiency syndrome; VRE, vancomycin-resistant enterococci. ‡ Includes other notifiable conditions. View Large From the 33 published studies, 56 measures of reporting completeness were recorded, involving 21 different infectious diseases reportable in the jurisdictions where the studies were conducted. Reporting completeness differed significantly among the 21 distinct disease categories (Kruskal-Wallis test, p < 0.05). No temporal trends in disease-reporting completeness were noted. Reporting completeness for studies conducted during the 1990s was similar to reporting completeness during the 1980s. Reporting completeness was generally lower for studies conducted during the 1970s, but these earlier studies preceded the identification of human immunodeficiency virus and mandatory reporting of AIDS. After stratification by disease, reporting completeness during the 1970s was similar to completeness during subsequent decades. No other factor (i.e., geographic location, study design, or study size), either individually or after stratification by disease, was significantly associated with reporting completeness. When we compared the individual measures of reporting completeness for tuberculosis, AIDS, and sexually transmitted diseases (n = 25) with those for all other diseases (n = 31), the mean of the individual measures of reporting completeness was significantly higher for tuberculosis, AIDS, and sexually transmitted diseases (79 percent) than for all other diseases (49 percent; t test, p < 0.01). DISCUSSION The published literature reveals variation in infectious disease-reporting completeness in the United States, ranging from 9 to 99 percent. Though highly variable, reporting completeness appears to be most strongly related to the disease or condition being reported. Other factors, such as geographic location, appear to be less important in determining disease-reporting completeness. No temporal trends were noted and, based on the studies we reviewed, infectious disease-reporting completeness seems neither to have improved nor to have deteriorated appreciably in the United States during the last three decades. Reporting completeness was significantly greater for tuberculosis, AIDS, and sexually transmitted diseases as a group than for all other notifiable infectious diseases combined. The reasons for this are not clear but may be related to the perceived seriousness of these diseases or to the greater financial and human resources devoted to treating and preventing them, often involving contact tracing by case workers in the community. Historically, the strong reliance of tuberculosis and sexually transmitted disease patients on the public health system for case management and treatment has also augmented the surveillance for these conditions. Studies conducted in other industrialized countries have found patterns of reporting completeness similar to those found in the United States (45–52). Both in the United States and abroad, AIDS surveillance has generally been more frequently evaluated and AIDS cases more completely reported than for other diseases. The only other published reviews on the subject of infectious disease-reporting completeness have focused exclusively on AIDS reporting (53, 54). Buehler et al. (53) reviewed several published and unpublished AIDS surveillance evaluation activities in the United States, whereas the review by Gertig et al. (54) provides an international comparison of AIDS-reporting completeness. Ours is the first known review of reporting completeness for infectious diseases in general, and it suggests important differences depending on the condition under surveillance. Studies in this review provide general support for the commonly held belief that active disease surveillance results in more complete case enumeration than passive surveillance methods. This observation has been frequently noted and further suggested by studies that were unable to control for true temporal or spatial differences in disease incidence (55–58). Evaluation activities that supplement routine data sources to derive a measure of reporting completeness are, albeit temporary, a form of active surveillance. Such evaluation activities, whether published or not, undoubtedly contribute to more complete national disease reporting (53). However, because national infectious disease incidence data are derived from a mixture of active and passive surveillance methods, the observation that active surveillance results in increased reporting completeness does not readily provide a clearer interpretation of national data. Thus, as a means to both better understand surveillance data and contribute to increased reporting completeness, ongoing surveillance evaluation of the kind represented by studies in this review should be encouraged. Previous reports have delineated the specific stages in the disease identification and reporting process for conditions of public health importance. Using shigellosis as an example, Rosenberg et al. (59) estimated that, for every 100 persons infected with Shigella, 76 became symptomatic, 28 consulted a physician, nine submitted stool cultures, seven had positive culture results, six were reported to the local health department, and five were reported nationally to the Centers for Disease Control and Prevention. Thus, they proposed a multiplication factor of 20 to estimate the overall number of Shigella infections based on national shigellosis case reports. Similar strategies have been used to estimate the number of Salmonella (60) and Chlamydia (61) infections in the United States. Our review of the literature specifically addressed the later stage in this process between clinical or laboratory diagnosis and reporting to local, state, or national health authorities. Early components of surveillance sensitivity related to underdiagnosis, such as the proportion of those infected who become symptomatic, seek treatment, submit diagnostic specimens, and obtain laboratory confirmation of infection, were not addressed by this review but are also important to consider when interpreting surveillance data. Several authors have described reasons for failure of health-care providers and laboratories to report notifiable diseases (9, 62–66). The reasons cited include a lack of awareness of the legal requirement to report, a lack of knowledge of which diseases are reportable, a lack of understanding of how or to whom to report, an assumption that someone else will report the case, intentional failure to report to protect patient privacy, and insufficient reward for reporting or penalty for not reporting. Interventions aimed at reducing these barriers have had limited success at improving provider and laboratory reporting behavior (9, 64–66). Advances in provider- and laboratory-based information management hold promise for automated disease reporting and surveillance. The use of electronic health information systems should help to reduce dependence on individual provider behavior for routine disease surveillance, yet still allow patient privacy and confidentiality to be maintained. One study in our review demonstrated the added completeness of disease reporting achievable through the use of automated systems for capturing electronic laboratory data (43). When compared with traditional paper-based morbidity reporting, automated reporting of electronic laboratory data was found by Effler et al. to have resulted in a 2.3-fold increase in case reports. Furthermore, electronic reports arrived an average of 3.8 days earlier than conventional reports and were more likely to provide complete patient-, specimen-, and provider-related information. Another study demonstrated the usefulness of automated health maintenance organization data for detecting unreported cases of tuberculosis (29). Yokoe et al. found that health maintenance organization pharmacy-dispensing data, suggesting the use of two or more antituberculosis drugs, were a useful marker for identifying tuberculosis patients, particularly those without positive microbiologic results. As further advances are made in information technology and clinical and laboratory data standards, coupled with further consolidation of health-care delivery in the marketplace, automated electronic provider- and laboratory-based reporting will likely become a more prominent component of routine disease surveillance in the United States (67). One third of the studies included in our review used capture-recapture methods to estimate the number of cases undetected by all available data sources. Such methods are based on the key assumption that the data sources are independent; that is, the probability of being present or “captured” in one data source is unrelated to the probability of being present in the other data sources (68). These methods can result in inaccurate estimates when patients are misdiagnosed or when patient diagnoses are inaccurately recorded in health information systems (69, 70). Therefore, data validation is an important aspect of accurately applying capture-recapture methods. Estimates using capture-recapture methods are also influenced by record linkage errors (71). False negative matches will underestimate reporting completeness, whereas false positive matches will inflate estimates of reporting completeness. Therefore, for these methods to be successfully applied, appropriate criteria must be defined for accurately matching persons between data sources. However, for any given study, applying underascertainment corrected methods will always result in an equal or lower estimation of reporting completeness than would result from using uncorrected methods. This is evidenced by Prevots et al. (35) when evaluating completeness of reporting for vaccine-associated paralytic polio. Their conclusion of 81 percent (92 of 114) reporting completeness using capture-recapture methods is lower than the 94 percent (92 of 98) completeness that could be confirmed using uncorrected methods. Whether underascertainment corrected methods result in a more accurate estimate of reporting completeness depends on how these methods are applied and the individual characteristics of the data sources being used (72, 73). Despite the effect of different methods on any given study in our review, study design was not significantly associated with reporting completeness. Therefore, studies using the two different methods were combined for our analyses. In conclusion, surveillance systems require ongoing maintenance and evaluation if the data that result from them are to be accurately interpreted. We have presented here a review of quantitative evaluations of infectious disease surveillance and case-reporting completeness conducted in the United States during the last three decades. As health-care services and information technology continue to evolve, the possibility exists for numerous changes in conducting routine public health surveillance. Ongoing evaluation of disease-reporting completeness will continue to be a necessary part of public health surveillance, enabling more accurate interpretation of surveillance information for disease control and prevention. Reprint requests to Timothy Doyle, Epidemiology Program Office, Centers for Disease Control and Prevention, Mail Stop K-74, 4770 Buford Highway, Atlanta, GA 30341 (email: TDoyle@cdc.gov). The authors thank Lori Hutwagner for statistical advice and Dr. G. David Williamson for critical review of the manuscript. REFERENCES 1. Roush S, Birkhead G, Koo D, et al. Mandatory reporting of diseases and conditions by health care professionals and laboratories. JAMA  1999; 282: 164–70. Google Scholar 2. Summary of notifiable diseases, United States, 1998. MMWR Morb Mortal Wkly Rep  1999; 47: ii–92. Google Scholar 3. Guidelines for evaluating surveillance systems. MMWR Morb Mortal Wkly Rep  1988; 37(suppl 5): 1–18. Google Scholar 4. Thacker SB, Choi K, Brachman PS. The surveillance of infectious diseases. JAMA  1983; 249: 1181–5. Google Scholar 5. Cates W, Williamson GD. Descriptive epidemiology: analyzing and interpreting surveillance data. In: Teutsch S, Churchill R, eds. Principles and practices of public health surveillance. New York, NY: Oxford University Press, 1994:96–135. Google Scholar 6. Fleming WL, Brown WJ, Donohue JF, et al. National survey of venereal disease treated by physicians in 1968. JAMA  1970; 211: 1827–30. Google Scholar 7. Meek JI, Roberts CL, Smith EV Jr, et al. Underreporting of Lyme disease by Connecticut physicians, 1992. J Public Health Manag Pract  1996; 2: 61–5. Google Scholar 8. Eisenberg MS, Wiesner PJ. Reporting and treating gonorrhea: results of a statewide survey in Alaska. J Am Vener Dis Assoc  1976; 3: 79–83. Google Scholar 9. Rothenberg R, Bross DC, Vernon TM. Reporting of gonorrhea by private physicians: a behavioral study. Am J Public Health  1980; 70: 983–6. Google Scholar 10. Stroup DF. Special analytic issues. In: Teutsch S, Churchill R, eds. Principles and practice of public health surveillance. New York, NY: Oxford University Press, 1994:136–49. Google Scholar 11. McCarty DJ, Tull ES, Moy CS, et al. Ascertainment corrected rates: applications of capture-recapture methods. Int J Epidemiol  1993; 22: 559–65. Google Scholar 12. Marier R. The reporting of communicable diseases. Am J Epidemiol  1977; 105: 587–90. Google Scholar 13. Murray RJ, Hayden CH, Zahn F. Irregular reporting of tuberculosis cases by laboratories in Nassau County, N.Y. Public Health Rep  1974; 89: 385–8. Google Scholar 14. Kimball AM, Thacker SB, Levy ME. Shigella surveillance in a large metropolitan area: assessment of a passive reporting system. Am J Public Health  1980; 70: 164–6. Google Scholar 15. Davis JP, Bohn MJ. The extent of under-reporting of meningococcal disease in Wisconsin: 1980–1982. Wis Med J  1984; 83: 11–14. Google Scholar 16. Vogt RL, Clark SW, Kappel S. Evaluation of the state surveillance system using hospital discharge diagnoses, 1982–1983. Am J Epidemiol  1986; 123: 197–8. Google Scholar 17. Chamberland ME, Allen JR, Monroe JM, et al. Acquired immunodeficiency syndrome in New York City: evaluation of an active surveillance system. JAMA  1985; 254: 383–7. Google Scholar 18. Harkess JR, Gildon BA, Archer PW, et al. Is passive surveillance always insensitive? An evaluation of shigellosis surveillance in Oklahoma. Am J Epidemiol  1988; 128: 878–81. Google Scholar 19. Lindan CP, Hearst N, Singleton JA, et al. Underreporting of minority AIDS deaths in San Francisco Bay area, 1985–86. Public Health Rep  1990; 105: 400–4. Google Scholar 20. Conway GA, Colley-Niemeyer B, Pursley C. Underreporting of AIDS cases in South Carolina, 1986 and 1987. JAMA  1989; 262: 2859–63. Google Scholar 21. Modesitt SK, Hulman S, Fleming D. Evaluation of active versus passive AIDS surveillance in Oregon. Am J Public Health  1990; 80: 463–4. Google Scholar 22. Campos-Outcalt D, England R, Porter B. Reporting of communicable diseases by university physicians. Public Health Rep  1991; 106: 579–83. Google Scholar 23. Fife D, MacGregor RR, McAnaney J. Limitations of AIDS reporting under favorable circumstances. Am J Prev Med  1993; 9: 317–20. Google Scholar 24. Kirsch TD, Shesser R, Barron M. Disease surveillance in the ED: factors leading to the underreporting of gonorrhea. Am J Emerg Med  1998; 16: 137–40. Google Scholar 25. Elcock M, Simon T, Gilbert BP, et al. Active AIDS surveillance: hospital-based case finding in a metropolitan California county. Am J Public Health  1993; 83: 1002–5. Google Scholar 26. Standaert SM, Lefkowitz LB Jr, Horan JM, et al. The reporting of communicable diseases: a controlled study of Neisseria meningitidis and Haemophilus influenzae infections. Clin Infect Dis  1995; 20: 30–6. Google Scholar 27. Ewert DP, Westman S, Frederick PD, et al. Measles reporting completeness during a community-wide epidemic in inner-city Los Angeles. Public Health Rep  1995; 110: 161–5. Google Scholar 28. Driver CR, Braden CR, Nieves RL, et al. Completeness of tuberculosis case reporting, San Juan and Caguas regions, Puerto Rico, 1992. Public Health Rep  1996; 111: 157–61. Google Scholar 29. Yokoe DS, Subramanyan GS, Nardell E, et al. Supplementing tuberculosis surveillance with automated data from health maintenance organizations. Emerg Infect Dis  1999; 5: 779–87. Google Scholar 30. Smucker DR, Thomas JC. Evidence of thorough reporting of sexually transmitted diseases in a southern rural county. Sex Transm Dis  1995; 22: 149–54. Google Scholar 31. Dembek ZF, Kellerman SE, Ganley L, et al. Reporting of vancomycin-resistant enterococci in Connecticut: implementation and validation of a state-based surveillance system. Infect Control Hosp Epidemiol  1999; 20: 671–5. Google Scholar 32. Trepka MJ, Beyer TO, Proctor ME, et al. An evaluation of the completeness of tuberculosis case reporting using hospital billing and laboratory data; Wisconsin, 1995. Ann Epidemiol  1999; 9: 419–23. Google Scholar 33. Cochi SL, Edmonds LE, Dyer K, et al. Congenital rubella syndrome in the United States, 1970–1985: on the verge of elimination. Am J Epidemiol  1989; 129: 349–61. Google Scholar 34. Sutter RW, Cochi SL, Brink EW, et al. Assessment of vital statistics and surveillance data for monitoring tetanus mortality, United States, 1979–1984. Am J Epidemiol  1990; 131: 132–42. Google Scholar 35. Prevots DR, Sutter RW, Strebel PM, et al. Completeness of reporting for paralytic poliomyelitis, United States, 1980 through 1991. Implications for estimating the risk of vaccine-associated disease. Arch Pediatr Adolesc Med  1994; 148: 479–85. Google Scholar 36. Hardy AM, Starcher ET, Morgan WM, et al. Review of death certificates to assess completeness of AIDS case reporting. Public Health Rep  1987; 102: 386–91. Google Scholar 37. Sutter RW, Cochi SL. Pertussis hospitalizations and mortality in the United States, 1985–1988: evaluation of the completeness of national reporting. JAMA  1992; 267: 386–91. Google Scholar 38. Rosenblum L, Buehler JW, Morgan WM, et al. The completeness of AIDS case reporting, 1988: a multisite collaborative surveillance project. Am J Public Health  1992; 82: 1495–9. Google Scholar 39. Davis SF, Stebel PM, Atkinson WL, et al. Reporting efficiency during a measles outbreak in New York City, 1991. Am J Public Health  1993; 83: 1011–15. Google Scholar 40. Ikeda RM, Birkhead GS, Flynn MK, et al. Use of multiple reporting sources for perinatal hepatitis B surveillance and follow-up. Am J Epidemiol  1995; 142: 765–70. Google Scholar 41. Ackman DM, Birkhead G, Flynn M. Assessment of surveillance for meningococcal disease in New York State, 1991. Am J Epidemiol  1996; 144: 78–82. Google Scholar 42. Barat LM, Barnett BJ, Smolinski MS, et al. Evaluation of malaria surveillance using retrospective, laboratory-based active case detection in four southwestern states, 1995. Am J Trop Med Hyg  1999; 60: 910–14. Google Scholar 43. Effler P, Ching-Lee M, Bogard A, et al. Statewide system of electronic notifiable disease reporting from clinical laboratories. JAMA  1999; 282: 1845–50. Google Scholar 44. Wittes JT, Colton T, Sidel VW. Capture-recapture methods for assessing the completeness of case ascertainment when using multiple information sources. J Chronic Dis  1974; 27: 25–36. Google Scholar 45. Calzavara LM, Coates RA, Craib KJ, et al. Underreporting of AIDS cases in Canada: a record linkage study. CMAJ  1990; 142: 36–9. Google Scholar 46. Johnson RJ, Montano BL, Wallace EM. Using death certificates to estimate the completeness of AIDS case reporting in Ontario in 1985–87. CMAJ  1989; 141: 537–40. Google Scholar 47. Rivest P, Sagot B, Bedard L. Evaluation of the completeness of reporting of invasive meningococcal disease. Can J Public Health  1999; 90: 250–2. Google Scholar 48. Clarkson JA, Fine PE. The efficiency of measles and pertussis notification in England and Wales. Int J Epidemiol  1985; 14: 153–68. Google Scholar 49. Macleod CA. Haemophilus influenzae: the efficiency of reporting invasive disease in England and Wales. Commun Dis Rep CDR Rev  1994; 4: R13–16. Google Scholar 50. Jenkinson D. Whooping cough: what proportion of cases is notified in an epidemic? Br Med J (Clin Res Ed)  1983; 287: 185–6. Google Scholar 51. Bernillon P, Lievre L, Pillonel J, et al. Record-linkage between two anonymous databases for a capture-recapture estimation of underreporting of AIDS cases: France 1990–1993. Int J Epidemiol  2000; 29: 168–74. Google Scholar 52. Rushworth RL, Bell SM, Rubin GL, et al. Improving surveillance of infectious diseases in New South Wales. Med J Aust  1991; 154: 828–31. Google Scholar 53. Buehler JW, Berkelman RL, Stehr-Green JK. The completeness of AIDS surveillance. J Acquir Immune Defic Syndr  1992; 5: 257–64. Google Scholar 54. Gertig DM, Marion SA, Schechter MT. Estimating the extent of underreporting in AIDS surveillance. AIDS  1991; 5: 1157–64. Google Scholar 55. Thacker SB, Redmond S, Rothenberg RB, et al. A controlled trial of disease surveillance strategies. Am J Prev Med  1986; 2: 345–50. Google Scholar 56. Vogt RL, LaRue D, Klaucke DN, et al. Comparison of an active and passive surveillance system of primary care providers for hepatitis, measles, rubella, and salmonellosis in Vermont. Am J Public Health  1983; 73: 795–7. Google Scholar 57. Alter MJ, Mares A, Hadler SC, et al. The effect of underreporting on the apparent incidence and epidemiology of acute viral hepatitis. Am J Epidemiol  1987; 125: 133–9. Google Scholar 58. Levy BS, Mature J, Washburn JW. Intensive hepatitis surveillance in Minnesota: methods and results. Am J Epidemiol  1977; 105: 127–34. Google Scholar 59. Rosenberg ML, Marr JS, Gangarosa EJ, et al. Shigella surveillance in the United States, 1975. J Infect Dis  1977; 136: 458–60. Google Scholar 60. Chalker RB, Blaser MJ. A review of human salmonellosis. III. Magnitude of Salmonella infection in the United States. Rev Infect Dis  1988; 10: 111–24. Google Scholar 61. Groseclose SL, Zaidi AA, DeLisle SJ, et al. Estimated incidence and prevalence of genital Chlamydia trachomatis infections in the United States, 1996. Sex Transm Dis  1999; 26: 339–44. Google Scholar 62. Cleere RL, Dougherty WJ, Fiumara NJ, et al. Physicians' attitudes toward venereal disease reporting. JAMA  1967; 202: 117–22. Google Scholar 63. Jones JL, Meyer P, Garrison C, et al. Physician and infection control practitioner HIV/AIDS reporting characteristics. Am J Public Health  1992; 82: 889–91. Google Scholar 64. Weiss BP, Strassburg MA, Fannin SL. Improving disease reporting in Los Angeles County: trial and results. Public Health Rep  1988; 103: 415–21. Google Scholar 65. Schramm MM, Vogt RL, Mamolen M. The surveillance of communicable disease in Vermont: who reports? Public Health Rep  1991; 106: 95–7. Google Scholar 66. Konowitz PM, Petrossian GA, Rose DN. The underreporting of disease and physicians' knowledge of reporting requirements. Public Health Rep  1984; 99: 31–5. Google Scholar 67. McDonald CJ, Overhage JM, Dexter P, et al. A framework for capturing clinical data sets from computerized sources. Ann Intern Med  1997; 127: 675–82. Google Scholar 68. Brenner H. Use and limitations of the capture-recapture method in disease monitoring with two dependent sources. Epidemiology  1994; 6: 42–8. Google Scholar 69. Brenner H. Effects of misdiagnoses on disease monitoring with capture-recapture methods. Clin Epidemiol  1996; 49: 1303–7. Google Scholar 70. Washko RM, Frieden TR. Tuberculosis surveillance using death certificate data, New York City, 1992. Public Health Rep  1996; 111: 251–5. Google Scholar 71. Brenner H. Application of capture-recapture methods for disease monitoring: potential effects of imperfect record linkage. Methods Inf Med  1994; 33: 502–6. Google Scholar 72. Hook EB, Regal RR. Capture-recapture methods in epidemiology: methods and limitations. Epidemiol Rev  1995; 17: 243–64. Google Scholar 73. Hook EB, Regal RR. Recommendations for presentation and evaluation of capture-recapture estimates in epidemiology. J Clin Epidemiol  1999; 52: 917–26. Google Scholar
TI - Completeness of Notifiable Infectious Disease Reporting in the United States: An Analytical Literature Review
JO - American Journal of Epidemiology
DO - 10.1093/aje/155.9.866
DA - 2002-05-01
UR - https://www.deepdyve.com/lp/oxford-university-press/completeness-of-notifiable-infectious-disease-reporting-in-the-united-e6PAMugJ73
SP - 866
EP - 874
VL - 155
IS - 9
DP - DeepDyve
ER -