Recent trends in the incidence and morbidity that are associated with perinatal human immunodeficiency virus infection in the United States

Article Outline

• Abstract
• Methods
  • Data source
  • Model structure for estimating infections by birth year
  • Statistical analysis
• Results
  • Trends in perinatal AIDS cases from 50 states and the District of Columbia
  • Perinatal HIV infection cases in 33 states with name-based reporting
• Comment
• Acknowledgment
• References
• Copyright

We analyzed national surveillance data that had been reported to the Centers for Disease Control and Prevention to elucidate the impact of recent clinical and public health efforts to further decrease the number of human immunodeficiency virus (HIV) infections and resulting morbidity caused by perinatal transmission. Long-term trends in pediatric (ages, 0-13 years), perinatal acquired immune deficiency syndrome (AIDS) cases were analyzed by log-linear Poisson regression for the period 1992-2004. Estimates for the number of perinatal HIV infections that occurred during the more recent period of 2001-2004 were developed by extrapolation from the 33 states with ongoing HIV (non-AIDS) reporting to the entire United States with the use of a probabilistic model. The number of pediatric perinatal AIDS cases that were identified decreased from 858 in 1992 to only 41 in 2004. These declines were consistent across demographic and regional subgroups. Data on the number of perinatal HIV infections suggests ongoing declines throughout the early years of the 21st century from 277 (95% CI, 224-346) in 2001 to 138 (95% CI, 96-186) in 2004. The incidence and morbidity associated with perinatal HIV infection continue to decline. To ensure that existing prevention efforts continue to achieve control of these infections, consistent methods of public health surveillance must be instituted throughout the entire United States.

Key words: HIV, perinatal, surveillance, United States

A cornerstone of prevention and public health practice is the ongoing, systematic collection, analysis, evaluation, and dissemination of data that describe and monitor health events such as disease incidence and prevalence and the determinants of these outcomes.¹ All data sources and formal evaluations to date suggest that the prevention of the transmission of human immunodeficiency virus (HIV) from mother to infant (perinatal transmission) in the United States has been very successful.², ³ This success has been achieved through the implementation of guidelines for the identification of HIV-infected women through testing, the provision of comprehensive preconception (including contraceptive services) and perinatal care, the administration of antiretroviral therapy to the mother during pregnancy and delivery, and the administration of antiretrovirals to the HIV-exposed newborn infant.⁴, ⁵ However, despite documented decreases in perinatally associated acquired immunodeficiency syndrome (AIDS) cases since the apex of this epidemic in 1992, enumeration of the number of infected infants has relied on indirect statistical techniques rather than systematic public health reports of such cases.⁶, ⁷, ⁸, ⁹ The most recent national estimate from the Centers for Disease Control and Prevention (CDC) was calculated for the year 2000 and indicated that between 280 and 370 HIV-infected infants were born that year.¹⁰ That report estimated that only about one third of these infections (80-110) could have been prevented through improved identification and treatment of pregnant HIV-infected women. The evaluation of the impact of increased efforts to identify such women through policies such as routine testing requires ongoing and timely estimates of the number of perinatal infections that occur in the areas that implement such policies and an assessment of the burden in terms of transmissions and morbidity because of these infections.², ¹¹, ¹²

In the past, the only comprehensive national data that are available on perinatal transmissions have been through the reporting of children with infections that progress to AIDS. However, with improved treatment of HIV infections, most AIDS cases can be prevented.¹³ Therefore, reporting that is based on AIDS cases no longer provides timely and accurate information about perinatal HIV transmission. Unfortunately, data on perinatally infected children have not been available at the national level because mandated reporting to state and local health departments of HIV-infected persons with disease that has not progressed to AIDS has been implemented inconsistently.¹⁴ Responsibility and legal mandates for disease reporting lies with state and local health departments, not the federal government; thus, reporting practices differ from state to state. The stigma that is associated with the HIV infection has engendered concerns that governmental registries that contain the names of HIV-infected persons might lead to discrimination and that apprehension about such reporting could be a deterrent to being able to recruit persons for HIV testing. As a result, many state health departments have found it politically difficult to establish such reporting systems for HIV infection.⁶ However, systematic surveys of persons who are at risk of HIV have demonstrated consistently that fear about disease reporting is not among the most commonly identified deterrents, and testing rates have not declined in areas that have implemented name-based reporting for this condition.¹⁴, ¹⁵ Therefore, throughout the late 1990s and into the 21st century, more states began to implement such data collection systems.

By 2001, 33 states had established name-based HIV-infection reporting systems that provide data that can be processed and analyzed by the CDC. These areas report 70% of the perinatally associated AIDS cases in the United States. The availability of data from areas with the most burden of perinatal HIV transmission provides an opportunity to use more direct methods to estimate the number of perinatally acquired infections occuring in the United States and to review recent trends in the morbidity that is associated with HIV.

Back to Article Outline

Methods 

Data source 

The data used for this analysis are from the national reporting system for HIV/AIDS. The analysis includes records of AIDS and HIV cases reported and maintained by the CDC through December of 2005. The details of the methods and operations that are used to compile the data that is entered into this system have been described in more detail elsewhere.¹⁶, ¹⁷ Briefly, all states and the District of Columbia have mandated that healthcare providers (including physicians, laboratories, hospital and clinic personnel, and in some cases insurance companies) submit case reports to appropriate local public health authorities on all persons who are diagnosed with AIDS with the use of a standard case definition. These reports include identification of information about the patient that includes the name and usually the address of the case. Data from the reports are then forwarded to the CDC after the personal-identifying information is removed from the case record.

In 1999 CDC published guidelines for the operation of HIV reporting systems that included criteria for the evaluation of the systems and a standardized case definition for HIV, non-AIDS infections in adults, adolescents, and children who are infected through perinatal exposure.¹⁸ However, unlike AIDS reporting for which local surveillance records included the patient’s name in all areas, many jurisdictions used identifiers other than the name of patients who were diagnosed with HIV infections that had not progressed to AIDS. Formal evaluations of these systems by the CDC indicated that the data from these systems were not sufficiently accurate to use for epidemiologic or other public health purposes.¹⁸, ¹⁹ Therefore, the CDC continues to exclude data for HIV non-AIDS cases from areas that use identifiers other than names in the national dataset.²⁰

As of 2001, 33 states had established name-based HIV reporting systems that included the submission of reports on children who were found to be infected through perinatal exposure to HIV. These states were Alabama, Alaska, Arkansas, Arizona, Colorado, Florida, Idaho, Indiana, Iowa, Kansas, Louisiana, Michigan, Minnesota, Mississippi, Missouri, Nebraska, Nevada, New Jersey, New Mexico, North Carolina, North Dakota, New York, Ohio, Oklahoma, South Carolina, South Dakota, Tennessee, Texas, Utah, Virginia, West Virginia, Wisconsin, and Wyoming. In the construction of a statistical model to estimate the total number of perinatal infections in the United States, data regarding children who were infected with HIV through perinatal exposure and in whom AIDS later developed are included from all states, and only data from the 33 areas with name-based reporting are included on cases that have not progressed to AIDS.

Model structure for estimating infections by birth year 

Table 1 shows the assumptions, data sources, and estimated uncertainty that are associated with each component of the model that was used to compute the national estimates. The fundamental assumption underlying this approach is that the number of new infections that were acquired in the areas with name-based reporting contributes to the national total by the same proportion as the number of childhood (age, <13 years) perinatal AIDS cases from those areas contributes to the total number of perinatally acquired AIDS cases in children who live in the entire United States. We used the proportion of all perinatal AIDS cases that were reported in the United States and that were reported from these 33 states over the 5-year period, 2001-2005. Therefore, the number of cases that were reported for a single birth year was multiplied by the inverse of the proportion of perinatal AIDS cases that were reported from the areas with name-based reporting. During this period, black children constituted 70% (Table 1) of case reports, even though only 12%-15% of all children who were born in the United States during this period were from this racial group. This disparity reflects the extraordinary racial disparity in HIV infections among women generally.²¹ Therefore, the estimates were developed for race-specific strata and aggregated to derive an overall national total.

TABLE 1. Variable values, sources for the values, and the probability distributions used in the model to estimate the number of infants infected with HIV in the United States by year, 2001-2005

Although a higher proportion of diagnosed HIV and AIDS cases are reported than for most other infectious diseases, there is ample information to demonstrate that not all cases are reported to health departments.²² In particular, the exclusion or confirmation of a diagnosis of perinatally acquired HIV in an exposed infant is very complex because such newborn infants usually possess maternal antibodies against HIV.¹⁸ Therefore, the confirmation of a diagnosis of perinatal infection can be delayed, and children who are infected may never be reported to the health department. We adjusted for delays in reporting by calculating the proportion of cases that were reported in each year after 2001 through 2005 among the infected children who were born in 2001. For years of birth subsequent to 2001, the number of cases who were born and reported as infected was inflated by the inverse of the proportion of the total number of cases that were reported for 2001 over a period of time equivalent to the period between the birth year and 2005. For example, according to the December 2005 CDC data, 94% of infections among children who were born in 2001 were reported by the end of 2004. Therefore, the number of observed cases who were born in 2002 was multiplied by the inverse of 0.94 to adjust for the anticipated reporting delay. Finally, some cases are never reported to the health department. There are no recent data on the completeness of reporting specifically for perinatal cases. However, formal evaluations to estimate underreporting for the overall number of newly diagnosed patients demonstrate that approximately 15% of HIV and/or AIDS cases would not be reported to the health department during a period similar to that of this analysis.¹⁷, ²³ Therefore, we further inflated all estimated case counts by the inverse of 0.85 to adjust for this undercounting phenomena (Table 1).

Statistical analysis 

AIDS case counts were available from all 50 states and the District of Columbia. These counts were compiled by year of diagnosis and were adjusted for reporting delay with standard methods that have been developed by the CDC.²⁴ AIDS case counts were not adjusted for underreporting. Trends were assessed after 1992 (the apex of the number of cases) by the calculation of regression coefficients with the use of log-linear Poisson regression that was based on the annual case counts from 1992-2004.²⁵ Differences in coefficients that were calculated for the diagnosis year were assessed with the standard errors for these coefficients from the models and the calculation of Z tests for pairwise comparisons or the logit test for homogeneity for categories with >2 groups (eg, race/ethnicity and region of the United States).²⁵, ²⁶ To estimate the annual percent of decline (APD) during this period the regression coefficients were transformed using the formula: (e^β − 1) × 100%, where β is the coefficient for the year of diagnosis.

The HIV infection computation model was constructed based on a probabilistic method used frequently in economic analyses.²⁷ Data and computations were entered and executed in Excel (Microsoft Corporation, Redmond, WA) spreadsheets. The uncertainty that was associated with each input was quantified with the use of a plausible probability distribution (eg, Poisson for count data or the binomial distribution for proportions) or by the identification from the literature of minimum, most likely, and maximum values for the input estimate and the assumption of a triangular distribution that was defined by these 3 values. See Table 1 for the details of each estimate entered into the model. The software package @RISK (version 4.5; Palisades Corporation, Ithaca, NY) was used to simulate 1000 results for each output estimate by sequentially drawing independent values from the input distributions and conducting the necessary calculations. Monte Carlo techniques were used to conduct the sampling. The distribution of the values for each output estimate reflects the overall uncertainty that is associated with the input values. Ninety-five percent confidence intervals (95% CI) were derived by the identification of the 2.5 and 97.5 percentile values for the distribution of computational results from the @RISK simulations.

Back to Article Outline

Results 

Trends in perinatal AIDS cases from 50 states and the District of Columbia 

The number of perinatal AIDS cases peaked in all 50 states and the District of Columbia in 1992.⁷ From 1992 through 2004, 4805 perinatal AIDS cases were diagnosed in children <13 years old. However, there was a 95% decrease from 1992, when 858 cases were diagnosed, through 2004 when only 41 cases were diagnosed (Figure A). This represented a 22% year over year decline for this metric (95% CI, 22-24).

• 
  • View Large Image
  • Download to PowerPoint

• FIGURE. 

  Trends in perinatal AIDS cases in the United States, 1992-2004

• United States AIDS cases in children <13 years old that are attributed to perinatal transmission by A, year, B, race/ethnicity, C, region, and D, gender from 1992-2004.

• McKenna. Recent trends in perinatal HIV in the United States. AJOG 2007.

Cases among black children constituted 67% of the total for entire period. There was little annual variation in this proportion (Figure B). Correspondingly, there was also little variation in the APD in any of the 3 major racial and ethnic groups (Figure B; black, 21 cases [95% CI, 18-24]; white, 23 cases [95% CI, 20-26]; Hispanic, 22 cases [95% CI, 19-25]). Among Asian and Pacific Islanders (data not shown in the Figure because of small numbers), there was also a significant decline (APD, 18; 95% CI, 5-26), even though only 33 cases were diagnosed over the entire period. Only 16 cases were diagnosed among Native American patients, which resulted in an imprecise assessment of the overall trend (APD, 7; 95% CI, 5-18).

Notable regional differences in trends existed in the United States (P < .05, for homogeneity). These regions are defined by the National Center for Health Statistics.⁴ In the south (APD, 20; 95% CI, 18-23) and northeast (APD, 24; 95% CI, 21-27) United States, comparable numbers of cases were diagnosed during the 13 years that were analyzed, and the declines were similar (Figure C). However, the western (APD, 19; 95% CI, 15-22) and midwestern (APD, 18; 95% CI, 13-22) United States had fewer cases and smaller relative declines in the case counts.

There were no notable differences in the magnitude or trends in case counts between male and female patients during this period (Figure D).

Perinatal HIV infection cases in 33 states with name-based reporting 

The number of HIV infection cases that was attributed to perinatal exposure in the 33 states that reported data to the CDC is displayed on the basis of the year of birth and the year the cases were reported to the surveillance programs (Table 2). The presence of maternal antibodies to HIV in the newborn infant complicates the diagnosis of perinatal HIV infections; unless repeated viral detection tests (eg, polymerase chain reaction for RNA or DNA) are positive, the current case definition requires as many as 18 months to diagnose infection definitively on the basis of the results of standard antibody tests.¹⁸ Therefore, a substantial number of infants who were born in 2001 were not reported until 2005 (ie, 10 cases). However, trends in the number of cases by year of birth can be inferred by the assessment of the diagonal cells from left to right. For example, 39 cases who were born in 2001 were reported during their year of birth, whereas there were only 25 such cases born in 2004. It is unlikely that this difference is attributable to increases in time delays for reporting during this period because there was an even greater decline in the number of cases that were reported the year after birth during the same period (74 cases in the 2001 birth cohort, but only 33 in the 2004 cohort).

TABLE 2. Number of perinatally infected infants by year of birth in the year of the report to the national database among 33 states with HIV infection report that have reported since 2001

Year of birth	Year of report					Number of infected infants in the birth cohort
	2001	2002	2003	2004	2005	n	N (95% CI)⁎
2001	39	74	31	8	10	162	277(224-346)
2002	—	29	50	21	9	109	204(161-276)
2003	—	—	25	34	27	86	167(127-224)
2004	—	—	—	25	33	58	138(96-186)

CI, confidence interval

† Data from 33 states with HIV infection that were reported were extrapolated to the entire United States (N). Estimates include adjustments for delays in reporting and underreporting of cases.

These birth cohort data were used to estimate the number of perinatally infected children who were born in the entire United States in each year from 2001 through 2004 (Table 2). Because of the small number of cases and the uncertainties that are associated with each of the estimation model inputs, there was substantial overlap in the confidence intervals for each year-after-year comparison. However, in general these data supported a continuing decline in the number of cases during the early part of the 21st century with a significant difference (Z test, 3.6; P < .01) between the estimate for 2001 and 2004 (Table 2).

Back to Article Outline

Comment 

These results, based on the most extensive data available to date regarding the direct population-based reporting of HIV infection and morbidity, suggest that the incidence of perinatally acquired HIV infection continues to decline in the United States. The trends in perinatally acquired AIDS are more difficult to interpret because the observed decreases could result from improved treatment of infected infants that forestalls the development of the sequelae that are associated with HIV. However, in conjunction with the data on infections from the growing number of areas such as New York State that have a history of substantial incidence and morbidity from this route of infection, it appears that the AIDS trends arise from both a decrease in infections and a potential success in slowing disease progression. There were statistically significant regional differences observed in the AIDS trends. However, the absolute differences in these geographic trends were modest (range, 18%-24% per year). Smaller declines occurred in areas with smaller numbers of cases. This may reflect a statistical “floor” whereby the relative rate of change decreases because the number of events becomes small and further reductions can be difficult.²⁸ The continued decline in annual perinatal HIV infections reflect a dynamic that was first observed in the mid 1990s when the CDC estimated that the annual number of children who had been born with HIV had fallen from a high of 1650 in 1991 to 480 in 1996.⁷ By 2000 the CDC estimated that the number decreased further to between 280 and 370 infections.¹⁰

The estimated decrease in perinatal HIV infections could arise through 1 or some combination of mechanisms. The first is a decrease in the number of births to HIV-infected women or better prevention efforts among HIV-infected women who become pregnant and deliver. The number of births could have declined either because of a decrease in birth rates among these women or a decrease in the population of infected women. The latter explanation is unlikely because survival among HIV-infected women continues to improve and despite evidence of recent modest declines in the number of new diagnoses in this group, the overall prevalence in women continued to increase during the early years of the 21st century.¹⁷, ²¹ This growing prevalence was identified by the CDC in 2000 as the major reason that the number of perinatal infections could be reduced only by another one-third, even if all women who delivered were provided optimal antiretroviral therapy. However, data from New York State, where essentially all newborn infants are tested for HIV by the health department, demonstrate that the number of births to infected women decreased by 60% from 1990 through 2003,²⁹ which suggests that at least in 1 high morbidity area declining birth rates among HIV-infected women is making a major contribution to the decrease in perinatal HIV cases.

The second possible explanation that perinatal HIV infections continue to decrease is improvement in the provision of appropriate antiretroviral therapy to HIV-infected women who become pregnant and deliver. Such treatment is contingent on the identification of these women before or during pregnancy through HIV testing. CDC surveillance data from 28 states with name-based reporting of infants who were born to HIV-infected women (ie, HIV exposure reporting) during the period 2001-2004 demonstrated that the proportion of infected mothers that was reported in these areas who were diagnosed before or at delivery was quite high at 93%.² This is consistent with data from 24 states that participated in the CDC Enhanced Perinatal Surveillance system in which, from 1999-2001, the proportion of infected mothers who were diagnosed during or before delivery was 94%, and approximately 80% of women received antiretroviral prenatally or during the intrapartum period.⁴

Establishing the specific determinants of the trends in perinatal incidence would require more comprehensive information about birth trends in HIV-infected women and the patterns of care in these women. Federal funding for anonymous antenatal seroprevalence surveys in the United States is prohibited currently by Congress.² Therefore, the only feasible source of information for the elucidation of the determinants of perinatal infections has to be derived from monitoring perinatally exposed newborn infants. The CDC currently supports 10 states and the cities of Chicago and Philadelphia and Puerto Rico to continue the Enhanced Perinatal Surveillance program. This system collects extensive information regarding treatment of the mother and infant before and after delivery (special Surveillance supplement).⁴ In addition, the CDC, the Council of State and Territorial Epidemiologists, and the American Academy of Pediatrics recommend that children who are born to HIV-infected mothers should be reported and followed for infection and potential long-term effects of exposure to HIV and antiretroviral.⁶, ¹⁸, ³⁰ However, because >90% of these children will not be infected, there are concerns about privacy that make the universal implementation of these recommendations unlikely.⁶

Without universal perinatal HIV exposure reporting, population surveillance of perinatally infected children will remain the fundamental metric for the definition of the status and success of efforts to prevent this route of transmission. Direct HIV infection reporting data from only 33 states could be used in this analysis. The imprecision of the case counts for more recent years was further exacerbated by the fact that definitive determination of the infection status for a newborn infant who was perinatally exposed to HIV can require up to 18 months of monitoring. Therefore, estimates of the number of newly infected newborn infants in any temporally proximate period must be adjusted for inevitable delays in diagnosis and reporting to public health officials. Further complicating this issue is the small number of these infections, which makes the statistical extrapolations dependent on a limited set of small counts. Hence, strong confidence that these declines reflect success in prevention will require observation for several more years. However, as of July of 2006, 45 states had legal authority to conduct name-based HIV infection surveillance, and the recent federal legislation reauthorizing the Ryan-White act essentially has required states to implement HIV reporting to receive this federal funding for treatment and care of HIV-infected patients.³¹ Therefore, the precision and validity of these numbers should improve rapidly over the next several years and provide critical inputs for focusing resources to sustain and augment prevention efforts that already have proved successful and ameliorate problems when they are identified.

Back to Article Outline

Acknowledgment 

The authors acknowledge the commitment, skill, and tireless work of the state and local HIV/AIDS Surveillance Coordinators who collect and vigorously protect the privacy and confidentiality of the data that were used in this analysis.

Back to Article Outline

References 

1. Centers for Disease Control and Prevention. Updated Guidelines for Evaluating Public Health Surveillance Systems. MMWR. 2001;50(RR13):1–35
   • View In Article
2. Centers for Disease Control and Prevention. Achievements in public health (Reduction in perinatal transmission of HIV infection: United States, 1985-2005). MMWR Morb Mortal Wkly Rep. 2006;55:592–597
   • View In Article
3. Wortley PM, Lindegren ML, Fleming PL. Successful implementation of perinatal HIV prevention guidelines: a multistate surveillance evaluation. MMWR Recomm Rep. 2001;50:17–28
   • View In Article
   • MEDLINE
4. Centers for Disease Control and Prevention. Enhanced perinatal surveillance: United States, 1999-2001. Atlanta: US Department of Health and Human Services, CDC; 2004;Accessed March 27, 2007. Available at: http://www.cdc.gov/hiv/stats/hasrsupp.htm
   • View In Article
5. US Department of Health and Human ServicesPublic Health Service. Recommendations for use of antiretroviral drugs in pregnant HIV-1–infected women for maternal health and interventions to reduce perinatal HIV-1 transmission in the United States. Washington (DC): US Department of Health and Human Services, Public Health Service; 2006;Accessed March 27, 2007. Available at http://aidsinfo.nih.gov/ContentFiles/PerinatalGL.pdf
   • View In Article
6. Committee on Pediatric AIDS. Surveillance of pediatric HIV infection. Pediatrics. 1998;101:315–319
   • View In Article
   • CrossRef
7. Lindegren ML, Byers RH, Thomas P, et al. Trends in perinatal transmission of HIV/AIDS in the United States. JAMA. 1999;282:531–538
   • View In Article
   • MEDLINE
   • CrossRef
8. Byers RH, Caldwell MB, Davis S, Gwinn M, Lindegren ML. Projection of AIDS and HIV incidence among children born infected with HIV. Stat Med. 1998;17:169–181
   • View In Article
   • MEDLINE
   • CrossRef
9. Davis SF, Byers RH, Lindegren ML, Caldwell MB, Karon JM, Gwinn M. Prevalence and incidence of vertically acquired HIV infection in the United States. JAMA. 1995;274:952–955
   • View In Article
   • MEDLINE
10. Office of Inspector General. Reducing obstetrician barriers to offering HIV testing. Washington (DC): US Department of Health and Human Services, Office of Inspector General; 2002;report no. OEI-05-01-00260. Accessed March 27, 2007. Available at http://oig.hhs.gov/oei/reports/oei-05-01-00260.pdf
    • View In Article
11. Centers for Disease Control and Prevention. HIV prevention: new strategies for a changing epidemic: United States, 2003. MMWR Morb Mortal Wkly Rep. 2003;52:329–332
    • View In Article
    • MEDLINE
12. Centers for Disease Control and Prevention. Revised recommendations for HIV screening of pregnant women. MMWR Morb Mortal Wkly Rep. 2001;50:59–86
    • View In Article
    • MEDLINE
13. Gona P, Van Dyke RB, Williams PL, et al. Incidence of opportunistic and other infections in HIV-infected children in the HAART era. JAMA. 2006;296:292–300
    • View In Article
    • CrossRef
14. Institute of Medicine. Measuring what matters: allocation, planning, and quality assessment for the Ryan White CARE Act. Washington (DC): The National Academies Press; 2003;
    • View In Article
15. Nakashima AK, Horsley R, Frey RL, Sweeney PA, Weber JT, Fleming PL. Effect of HIV reporting by name on use of HIV testing in publicly funded counseling and testing programs. JAMA. 1998;280:1421–1426
    • View In Article
    • MEDLINE
    • CrossRef
16. Nakashima AK, Fleming PL. HIV/AIDS surveillance in the United States, 1981-2001. J Acquir Immune Defic Syndr. 2003;32(suppl):S68–S85
    • View In Article
    • CrossRef
17. Centers for Disease Control and Prevention. HIV/AIDS surveillance report, 2004. Atlanta: US Department of Health and Human Services, CDC; 2004;Accessed March 27, 2007. Available at: http://www.cdc.gov/hiv/topics/surveillance/resources/reports/past.htm#surveillance
    • View In Article
18. Centers for Disease Control and Prevention. Guidelines for national human immunodeficiency virus case surveillance, including monitoring for human immunodeficiency virus infection and acquired immunodeficiency syndrome. MMWR 1999;48(No. RR-13). Accessed March 27, 2007. Available at http://www.cdc.gov/mmwr/preview/mmwrhtml/rr4813a1.htm.
    • View In Article
19. Borchardt SM, Rawls AW, Benbow N, Dworkin MS. An evaluation of the patient code number for HIV case reporting. Public Health Rep. 2006;121:360
    • View In Article
    • MEDLINE
20. Gerberding JL, Jaffe HW. Dear colleague letter, April 22. Atlanta: US Department of Health and Human Services, CDC; 2003;Accessed March 27, 2007. Available at http://www.cdc.gov/hiv/projects/perinatal/2003/letter.htm
    • View In Article
21. McDavid K, Li J, Lee LM. Racial and ethnic disparities in HIV diagnoses for women in the United States. J Acquir Immune Defic Syndr. 2006;42:101–107
    • View In Article
    • MEDLINE
    • CrossRef
22. Doyle TJ, Glynn MK, Groseclose SL. Completeness of notifiable infectious disease reporting in the United States: an analytical literature review. Am J Epidemiol. 2002;155:866–874
    • View In Article
    • MEDLINE
    • CrossRef
23. Hall HI, Song R, Gerstle JE, Lee LM HIV/AIDS Reporting System Evaluation Group. Assessing the completeness of reporting of human immunodeficiency virus diagnoses in 2002-2003: capture-recapture methods. Am J Epidemiol. 2006;164:391–397
    • View In Article
    • MEDLINE
    • CrossRef
24. Green TA. Using surveillance data to monitor trends in the AIDS epidemic. Stat Med. 1998;17:143–154
    • View In Article
    • MEDLINE
    • CrossRef
25. Kleinbaum DG, Kupper LL, Muller KE. Applied regression analysis and other multivariable methods. In: 2nd ed. Boston: PWS-Kent; 1988;
    • View In Article
26. Woolf B. On estimating the relation between blood group and disease. Ann Hum Genet. 1955;19:251–253
    • View In Article
    • MEDLINE
    • CrossRef
27. Baltussen RM, Hutubessy RC, Evans DB, Murray CJ. Uncertainty in cost-effectiveness analysis: probabilistic uncertainty analysis and stochastic league tables. Int J Technol Assess Health Care. 2002;18:112–119
    • View In Article
    • MEDLINE
28. Greenland S. Dose-response and trend analysis in epidemiology: alternatives to categorical analysis. Epidemiology. 1995;6:356–365
    • View In Article
    • MEDLINE
    • CrossRef
29. New York City Department of Health and Mental Hygiene. Pediatric and adolescent HIV/AIDS: surveillance update, New York City; 2005. Accessed March 27, 2007. Available at: http://www.nyc.gov/html/doh/downloads/pdf/dires/ped12-2005.pdf.
    • View In Article
30. Council of State and Territorial Epidemiologists. Surveillance for perinatal HIV exposure. 2002;Atlanta, Ga; Position Statement 02-ID-04. Accessed March 27, 2007. Available at: http://www.cste.org/position%20statements/02-ID-04.pdf.
    • View In Article
31. Glynn MK, Lee LM, McKenna MT. The status of national HIV case surveillance, United States 2006. Pub Health Rep. 2007;122(suppl):63–71
    • View In Article