Abstract
Objectives
Male circumcision is linked to a reduction in the risk of HIV infection, sexually transmitted infections, penile inflammatory skin disorders, cancers, urinary tract infections, and other complications. We examined the extent to which the change in circumcision recommendation by the American Academy of Pediatrics in 1999 and Medicaid coverage status in states affected the total number of procedures performed.
Methods
We used data from the Nationwide Inpatient Sample for 1998-2011 collected annually by the Healthcare Cost and Utilization Project. We examined data on all male births in the United States with Medicaid and private health insurance. We then categorized births into 4 groups: (1) births with newborn male circumcision procedure, (2) births with Medicaid or private health insurance, (3) births that occurred in states where Medicaid coverage for newborn male circumcision was removed, and (4) births that occurred before or after the policy change. We used multivariable logistic regression to estimate the adjusted odds of newborn male circumcision.
Results
In the 10 states where a change in Medicaid policy occurred, circumcision frequency had a mean percentage-point decrease of 21.4% among Medicaid beneficiaries and 3.2% among private health insurance beneficiaries from before to after the policy change. In states where coverage was maintained, the change in circumcision frequency was negligible for Medicaid and private health insurance beneficiaries. These changes resulted in an estimated 163 456 potential circumcisions not performed.
Conclusion
Decreases in newborn male circumcision frequency correlated with the Medicaid policy change for the procedure. Efforts should be made to reduce barriers for cost-effective preventive procedures that promote health, such as newborn male circumcision.
Citing a lack of evidence supporting the benefits of newborn male circumcision, the American Academy of Pediatrics weakened its stance on the procedure in 1999, taking a more neutral position. 1 In the decade that followed, 10 of 50 US states dropped Medicaid coverage for the procedure. 2 However, in 2012, the American Academy of Pediatrics changed its position on male circumcision in light of new evidence, stating that the benefits of the procedure (eg, prevention of urinary tract infections, transmission of some sexually transmitted infections, and penile cancer) outweigh the risks of the procedure (eg, inflammation, bleeding, infection); however, a concomitant restoration of coverage for the procedure has not occurred. 3 The proportion of male newborns circumcised nationwide from 2000 to 2010 dropped by 4.4% (from 61.3% to 56.9%). 2 One cross-sectional study showed that hospitals in states that covered newborn male circumcision in 2004 had a 24% higher rate of newborns circumcised than hospitals in states that did not cover this procedure. 4 Those changes to Medicaid policies are likely to increase disparities in reproductive health, because the populations that are affected most by the changes in policies are those that are low income and therefore most vulnerable to health complications.
Evidence suggests that male circumcision is an effective preventive measure for reducing the transmission of a number of infectious diseases, including HIV infection from women to men by up to 60%-73% 5,6 and human papillomavirus (a 3.5-fold risk reduction), a major risk factor for cervical cancer in women. 7 Use of the procedure is also justified for the potential reduction of other sexually transmitted diseases and inflammatory conditions of the penis. 8 In some conditions, such as balanoposthitis, the disease occurs exclusively in uncircumcised males and can occur at any age. 9 -11 The prevention of these lesions can reduce complications such as phimosis and paraphimosis, 9 which has been shown to be a strong predictor in the development of invasive penile cancer. 12
Circumcision may result in infection, bleeding, and inflammation 13 ; however, the benefits of the procedure have been determined to outweigh the risks. 3 In addition, HIV prevalence in the United States is high among men who have sex with men, and no evidence supports circumcision as a preventive measure for this group. However, heterosexual acquisition of HIV accounts for nearly one-third of all new infections, 14 in which case circumcision is an effective preventive measure among males. 5
Given the health benefits of male circumcision and the removal of coverage for newborn male circumcision by Medicaid in 18 states since 1982, our study aimed to evaluate the changes in newborn circumcision frequency nationally and by state from 1998 to 2011. We also examined trends between health insurance type and newborn male circumcision frequency, before and after Medicaid removed coverage for the procedure. We predicted that people with private health insurance would have higher frequencies of newborn male circumcision than people with Medicaid in states that maintained and removed Medicaid coverage for newborn circumcision. We also predicted that in states that removed coverage for newborn male circumcision, the frequency of circumcision would decrease only for births covered by Medicaid. Among states that maintained constant coverage, we predicted no significant change in circumcision frequency among people with private health insurance or Medicaid. Thus, we expected changes in health insurance coverage to be the greatest contributor to the decline in national circumcision frequency observed from 1998 to 2011.
Methods
We obtained data for 1998-2011 from the Nationwide Inpatient Sample (NIS), a nationally representative sample of 20% of all public hospital inpatient visits. 15 We first restricted our study population to all male in-hospital births using the International Classification of Diseases, Ninth Revision, Clinical Modification (ICD-9-CM) diagnosis codes V30-39 and identified births with a record of a circumcision procedure (ICD-9-CM code 64.0). 16 Individual-level covariates included patient race/ethnicity (non-Hispanic Black, non-Hispanic White, Hispanic, non-Hispanic other), zip code income quartile (≤25%, 26%-50%, 51%-75%, ≥76%), 17 and health insurance type (Medicaid, private). 15 We considered records eligible for inclusion only if the primary form of health insurance was private or Medicaid. Hospital-level variables included state and region (Northeast, South, Midwest, West). We further categorized all births according to whether they occurred in a state that had made a change in Medicaid coverage at any point during the study period and whether the birth was before or after that change. 18
We first estimated the national circumcision frequency for each year in the study period. We then compared circumcision frequencies for births occurring in the index group (covered by Medicaid after a change in coverage policy went into effect) with circumcision frequencies for births with the reference conditions (ie, covered by private health insurance, in states that did not undergo a change in policy, before a change in coverage policy went into effect). To capture the net change, we averaged circumcision frequencies during the 5-year period before and after policy changes took effect in the 10 states that removed coverage for the procedure (2002: Arizona, Missouri, North Carolina; 2003: Florida, Montana, Utah; 2004: Maine; 2005: Idaho, Louisiana, Minnesota). To adjust for broad national secular trends in circumcision frequency in states that did not change coverage status, we divided the study into 2 periods: early (1998-2002) or late (2003-2011). For the 10 states that dropped Medicaid coverage for the procedure from 2002 through 2005, we omitted the year of the change to allow a 1-year washout period for the change to take effect. In subsequent analysis, we used multivariable logistic regression to estimate the odds ratios of newborn male circumcision, adjusting for race/ethnicity, zip code income quartile, and hospital region. Because the NIS does not recommend using the data to make state-level inferences because of a lack of representativeness at that level, we did not report any state-level estimates and focused our analysis on observations aggregated across all states with similar policies. 19 We conducted all analyses using SAS version 9.4 (SAS Institute Inc). The study did not require institutional review board approval because it did not require human participation and relied solely on publicly available de-identified data. Data on race were missing from 24% of the sample.
Results
We identified 6 032 014 newborn births from 1998 through 2011. We observed a 0.48% (95% CI, 0.47-0.49) average annual decline in the frequency of male circumcision (Figure 1). When stratified by race/ethnicity, non-Hispanic White and non-Hispanic Black newborns had similar frequencies of newborn male circumcision during the study period: 68.9% (95% CI, 68.9%-68.9%) and 66.7% (95% CI, 66.6%-66.8%), respectively. Hispanic newborns had the lowest male circumcision frequency (18.4%; 95% CI, 18.3%-18.5%).
Newborn male circumcision frequency, United States, 1998-2011 (N = 6 032 014). The dotted line indicates all newborn males who were circumcised in the United States. Data source: Nationwide Inpatient Sample. 14
Frequencies of newborn male circumcision were lower among people covered by Medicaid than among people with private health insurance. When we stratified by health insurance type during 1998-2011, newborns who were covered through private health insurance had an overall circumcision frequency of 66.9% (95% CI, 66.8%-67.0%), whereas newborns who were covered through Medicaid had a circumcision frequency of 44.4% (95% CI, 44.3%-44.5%). When we compared data on newborns with different types of coverage over time in states that removed coverage for the procedure with states that did not, we found little to no change for those states that retained coverage (59.9% [95% CI, 59.8%-60.0%] to 57.6% [95% CI, 57.5%-57.7%]), whereas circumcision frequencies in states that had removed coverage for circumcision declined by 14.4 percentage points, from 61.0% (95% CI, 60.9%-61.1%) to 46.6% (95% CI, 46.5%-46.7%; Figure 2A).
(A) Circumcision frequencies for Medicaid and privately insured newborns (N = 5 490 432), before (1998-2002) and after (2003-2011) removal of Medicaid coverage for the procedure in 10 US states (2002: Arizona, Missouri, North Carolina; 2003: Florida, Montana, Utah; 2004: Maine; 2005: Idaho, Louisiana, Minnesota). Data source: Nationwide Inpatient Sample. 14 (B) Newborn male circumcision frequencies by time, coverage, and health insurance type, before (1998-2002) and after (2003-2011) removal of Medicaid coverage for the procedure in 10 US states (2002: Arizona, Missouri, North Carolina; 2003: Florida, Montana, Utah; 2004: Maine; 2005: Idaho, Louisiana, Minnesota; N = 5 490 432). Data source: Nationwide Inpatient Sample. 14
When we stratified the change in newborn male circumcision frequency on births in states that dropped coverage and by type of health insurance, only frequencies in the Medicaid group in a state that removed coverage declined. Newborn male circumcision frequency declined more among Medicaid births than among private health insurance births. When we examined the postpolicy vs prepolicy change in newborn male circumcision frequencies in the subgroup of births covered by Medicaid, we found that circumcision dropped by 21.4 percentage points (95% CI, 21.3-21.9 percentage points), from 49.7% to 28.3%, after the policy change (P < .001; Figure 2B). That decline in frequency translated into a net difference of 163 456 circumcision procedures that were not performed among a total of 756 744 Medicaid births that occurred in a 5-year period after the policy change in states that removed coverage. Meanwhile, among newborns in states that retained coverage and had Medicaid, circumcision frequency increased from 46.3% to 46.8% during the same period. Among newborns with private health insurance, circumcision frequency decreased by 3.2% percentage points (95% CI, 3.1%-3.3%), from 67.5% (95% CI, 67.4%-67.6%) to 64.3% (95% CI, 64.2%-64.4%), in states where coverage was removed and by 1.7 percentage points (95% CI, 1.6%-1.8%), from 68.2% (95% CI, 68.1%-68.3%) to 66.5% (95% CI, 66.34%-66.6%), in states where coverage was not removed.
In the multivariable logistic regression model (Table), newborn male circumcision was associated with Medicaid coverage policy change (odds ratio = 0.46; 95% CI, 0.45-0.47), adjusted for race/ethnicity, zip code income quartile, and hospital region.
Logistic regression modeling of the probability of newborn male circumcision among newborns, United States, 1999-2011a
aData source: Nationwide Inpatient Sample. 14
bOdds of newborn male circumcision among Medicaid-covered births in states that removed Medicaid coverage for the procedure, adjusted for race/ethnicity, zip code income quartile, and hospital region.
Discussion
We found that the national change in newborn male circumcision frequency was likely associated with changes in health insurance coverage. Our findings are supported by previous studies that showed the relationship between circumcision frequency in states with and without Medicaid coverage and the number of procedures covered by Medicaid and private health insurers. 2,4 Our work extends those findings by examining the changes over time. As a result, we showed the relationship between states that removed coverage for the procedure and the effect it had on frequencies of circumcision among newborns covered by Medicaid and private health insurance, as well as the national trend.
Leading up to the American Academy of Pediatrics statement on newborn male circumcision in 1999, 1 circumcision frequency had been increasing, with a peak observed of 62.4% in 1999. After the American Academy of Pediatrics revised its position, frequency began to decrease; the lowest frequency (54.0%) was in 2009. In states where Medicaid removed coverage for newborn male circumcision, frequency of circumcision decreased significantly among people with Medicaid, because their primary health insurance type after coverage was removed. This finding suggests that newborns covered by Medicaid are being circumcised at lower frequencies and are more vulnerable to the effect of policy changes to preventive care than newborns with private health insurance. The concern behind our findings is that reducing access to circumcision puts low-income people, who are already at a greater risk of acquiring infections such as HIV than those of higher socioeconomic status, at an even greater risk. 20 When compared with states in which Medicaid coverage for the procedure was maintained, circumcision frequency increased slightly among newborns covered by Medicaid and decreased slightly among newborns with private health insurance.
Furthermore, we examined whether the change in Medicaid policy, which resulted in a decrease of approximately 21.4 percentage points of newborns circumcised, was sufficient to result in the nationwide decrease. The slight decline over time for births covered by private health insurance in states that dropped Medicaid coverage rules out the possibility that births in which parents desired to have the procedure performed simply switched to private health insurance rather than using Medicaid if they had the option. It is possible that the greater decline in circumcision frequency among privately insured newborns in states that removed coverage for the procedure compared with states that did not remove coverage for the procedure could have been the result of private health insurers similarly updating their own policies after the change to Medicaid. Although a causal relationship cannot be inferred, the finding that circumcision frequency did not change substantially in states that maintained coverage yet dropped by 14.4 percentage points in states that removed coverage suggests that the policy change should be recognized as the likeliest contributing factor to the overall decrease in circumcision frequency. In addition, the univariate decline in newborn male circumcision frequency from 49.7% to 28.3% within the Medicaid group after the policy change demonstrates that the decline is not likely to be attributable to differences in race/ethnicity, the economic composition of communities, or regional variation.
Our analysis also shows that income level does not directly contribute to the decrease in the newborn male circumcision frequency in the same way that health insurance type does. This relationship is supported by our observation of a similar decrease in circumcision frequency among various income groups after the removal of circumcision coverage by Medicaid, whereas the removal of coverage resulted in a decrease in circumcision frequency for people covered by Medicaid and not private health insurance. Because all income levels had similar decreases in newborn circumcision frequency, this finding may suggest that people who desire to have their child circumcised will do so regardless of income. However, it may also suggest that the choice people with Medicaid must make about whether their child receives the procedure depends on whether or not Medicaid covers the procedure. 4
Our results are crucial in shedding light on health disparities in the United States. The analysis by Warner et al 2 confirms our finding that people with private health insurance have a higher newborn male circumcision frequency than people covered through Medicaid. In some cases, the circumcision frequency was 22.5 percentage points higher among people with private health insurance than among people covered by Medicaid (66.9% vs 44.5%) and as much as 39.2 percentage points higher among people with private health insurance than among people with Medicaid (67.5% vs 28.3%) in states that do not cover the procedure. These results draw attention to the downstream effects of policy changes on populations and their disproportionate effect on children from families with low socioeconomic status when mediated by health insurance. If people covered by Medicaid were given full access to the procedure, circumcision frequency among newborns covered by Medicaid may be similar to circumcision frequency among newborns with private health insurance. Because male circumcision may be associated with a reduction in many health complications, people covered through Medicaid not only have a higher likelihood of acquiring certain preventable diseases than people with private health insurance, but people in states where Medicaid no longer supports coverage for the procedure are also at the greatest risk of experiencing such illnesses. Thus, the burden of disease falls heavily on children residing in states that had policy changes, which contributes to increases in health disparities that they will experience in the future. 21 One study that examined the future prevalence of disease in the United States found that if circumcision frequencies were to fall to the level observed in Western Europe, the lifetime prevalence of HIV infection, human papillomavirus infection, herpes simplex virus type 2 infection, and urinary tract infections would all increase. 22 Although other factors may explain the change in national circumcision frequency over time, our analysis provides evidence that the state-level Medicaid policy change is at least partly responsible for the overall decline.
Limitations
Our study had several limitations. First, the NIS data, although nationally representative, are not designed to be representative at the state level. Therefore, we refrained from making state-level inferences and instead estimated the averages across all states that removed coverage. Second, 24% of data on race/ethnicity was missing. Because we know that some racial/ethnic minority populations have lower frequencies of circumcision than White people, the decline in overall circumcision frequency may be the result of a shift in births toward populations that do not circumcise their children. Third, our study excluded data on circumcisions that were not performed in the hospital setting and those that were not captured in the NIS data set. In addition, our analysis excluded data on births without Medicaid or private health insurance coverage. As such, we predicted that the overall circumcision frequency was slightly higher than our reported results.
Lastly, although the number of states included in the NIS changes from year to year, we corrected for this analytically by cross-referencing the list of states that removed coverage with the NIS state data list to omit any states that were not included in the NIS during the pre–post years in which the coverage change occurred (ie, Idaho, Louisiana, Montana, and North Carolina).
Conclusion
Newborn male circumcision is an elective procedure with religious and cultural implications. Although health may be only one factor in a family’s decision to circumcise, efforts should be made to reduce barriers for cost-effective preventive procedures such as newborn male circumcision. Here, we demonstrate the sensitivity of Medicaid policy to expert recommendations, the tendency for health insurers to seek short-term cost-saving solutions by limiting access to controversial but proven beneficial procedures, and the effect of policy changes on interventions that prevent disease. Future areas of study should focus on understanding the trends associated with the infectious diseases and cancers outlined in our study in the states that removed Medicaid coverage for newborn male circumcision.
Footnotes
Declaration of Conflicting Interests
The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding
The author(s) received no financial support for the research, authorship, and/or publication of this article.
