Design and statistical analysis of oral medicine studies: common pitfalls.

A growing number of articles are emerging in the medical and statistics literature that describe epidemiologic and statistical flaws of research studies. Many examples of these deficiencies are encountered in the oral, craniofacial, and dental literature. However, only a handful of methodologic articles have been published in the oral literature warning investigators of potential errors that may arise early in the study and that can irreparably bias the final results. In this study, we briefly review some of the most common pitfalls that our team of epidemiologists and statisticians has identified during the review of submitted or published manuscripts and research grant applications. We use practical examples from the oral medicine and dental literature to illustrate potential shortcomings in the design and analysis of research studies, and how these deficiencies may affect the results and their interpretation. A good study design is essential, because errors in the analysis can be corrected if the design was sound, but flaws in study design can lead to data that are not salvageable. We recommend consultation with an epidemiologist or a statistician during the planning phase of a research study to optimize study efficiency, minimize potential sources of bias, and document the analytic plan.

Firearm storage practices of officers in a law enforcement agency in the South.

BACKGROUND: Law enforcement officers play an important role in promoting firearm safety. This study examined their firearm-related attitudes and practices. METHODS: We conducted an anonymous, self-administered survey of law enforcement officers in an agency in the South concerning firearm ownership, storage practices, and opinions. RESULTS: The 207 respondents (response rate=71%) were primarily white (60%) and male (89%). The proportions of respondents with (55%) and without (45%) children were similar. Eighty percent of the officers had firearms in addition to the one they were issued for work. Most stored firearms unlocked (59%) and loaded (68%); almost half (44%) reported storing firearms both unlocked and loaded. Eighty-five percent indicated that they felt an added need to protect themselves and family because of their job. Those who reported having firearms for self-protection were less likely than those who had firearms for recreation to store firearms securely. As compared to those without children, law enforcement officers with children were more likely to store firearms safely, and were especially likely to store firearms locked up (chi(2)=12.72, p<0.0001). Respondents favored background checks, mandatory safety training, and enforcement of storage laws. Three quarters approved of government safety regulations for handguns. CONCLUSIONS: Despite the law enforcement officers' prominent role in firearm safety promotion and support for initiatives that limit unauthorized access, our study found that many do not practice safe storage. Apparently, many officers keep their firearms stored unlocked and loaded for the purpose of protecting themselves and their families. Addressing concerns about personal safety is a necessary step in promoting safe storage to law enforcement officers.

Driver's license and voter registration lists as population-based sampling frames for rural African Americans.

PURPOSE: To compare coverage of a state driver's license list and county voter registration lists as frames for sampling rural African Americans. METHODS: Name, birth date, sex, and county were extracted from records for all 18--59 year-old African Americans residing in eight rural North Carolina counties and listed in the North Carolina Driver license file [obtained as a SAS dataset from the University of North Carolina (UNC) Highway Safety Research Center] and in machine-readable text files of registered voters (obtained from county boards of elections). Files were edited, merged, and matched by last name and date of birth to produce three files for each five-year age group, sex, and county: 1) persons listed only in the driver's file; 2) persons listed only in the voter's file; and 3) persons listed in both. RESULTS: The median percentages of unique persons found only on the driver's list, only on the voter's list, and on both lists were 54%, 17%, and 30% men, and 35%, 25%, and 41% women, respectively. CONCLUSIONS: The driver's list provided better coverage than did voter registration lists. Federal legislation that prohibits states from releasing driver's license lists for use in surveys removes a valuable resource for population-based research.

Retrospective vs. prospective coital frequency and menstrual cycle length in a contraceptive effectiveness trial.

PURPOSE: To determine how well information at enrollment would predict coital frequency and menstrual segment length during a prospective contraceptive effectiveness trial. METHODS: We compared retrospective reports of monthly coital frequency and menstrual segment (cycle) length with prospective information for women participating in a contraceptive trial of the Reality((R)) female condom. RESULTS: Participants reported slightly higher mean monthly coital frequency and slightly longer menstrual segments prior to the study than during the study (12.6 acts vs. 12.0 acts per month and 30.8 days vs. 28.4 days). We examined the actual distribution of differences between the retrospective and prospective measures and found considerable variability. Among the 195 participants studied, parous women were less likely to predict accurately menstrual segment length (OR 0.4; 95% CI 0.3--0.9), while older women were more likely to predict accurately coital frequency (OR 1.9; 95% CI 0.9--3.3). CONCLUSIONS: Coital frequency and menstrual segment length vary considerably over time. Hence, prospective data collection is essential to accurately characterize these variations and to properly interpret results from contraceptive trials and other studies concerned with fecundability and hazards of reproduction.

"Love our kids, lock your guns": a community-based firearm safety counseling and gun lock distribution program.

BACKGROUND: Safer storage practices may reduce injury rates by limiting youth access to firearms. OBJECTIVE: To determine if a firearm safety counseling and gun lock distribution program improved storage practices. DESIGN: Community-based before-after trial. SETTING: Urban county in central North Carolina. PARTICIPANTS: One hundred twelve adult gun owners recruited through a mass media advertising campaign. INTERVENTION: In the parking lot of a shopping mall, participants completed a survey, and were then provided with tailored counseling, gun safety information, a gun lock, and instructions to use it. MAIN OUTCOME MEASURES: Firearm storage practices, assessed by survey and personal interview (baseline) and telephone interview (6-month follow-up). RESULTS: Most participants were white (62%), men (63%), had children (58%), and owned a gun for protection (74%). At follow-up, of the 82 participants, 63 (77%) (up from 39 [48%]) reported storing their gun(s) in a locked compartment (P =.004), 59 (72%) (up from 0) reported using gun locks (P =.001), 61 (74%) (up from 57 [69%]) reported storing their ammunition locked in a separate location, 59 (72%) (up from 52 [63%]) reported storing their gun(s) unloaded, and 6 (7%) (down from 15 [18%]) reported storing firearms unlocked and loaded. Participants with children were more likely at baseline to store weapons unlocked and loaded (38 [59%] vs 19 [41%]; P =.02) but were more likely after counseling to lock their weapons (29 [58%] vs 14 [44%]) and remove guns from the home (5 [10%] vs 0 [0%]). CONCLUSIONS: This program prompted reporting of safer firearm storage practices, particularly among parents. Longer follow-up, verification of self-reports and correct use, testing of gun locks, and monitoring firearm injury rates after distribution programs are needed to establish the public health potential of this approach.

Risk assessment for gonococcal and chlamydial infections in young children undergoing evaluation for sexual abuse.

OBJECTIVE: Testing for gonorrhea (GC) and chlamydial (Ct) infection in children who are being evaluated for sexual abuse is invasive and costly. We developed selective criteria to limit unnecessary testing for these infections. METHODS: Over a 10-year period (May 1988 to May 1998), clinical information was collected in a prospectively designed database for all children ages 0 to 12 years by the WakeMed Child Sexual Abuse Team in Raleigh, North Carolina. The study population comprised 3040 (2414 girls and 626 boys) of the 3064 children evaluated for sexual abuse. Children were interviewed, examined, and tested by culture for GC and Ct orally, rectally, and genitally. Information from referral sources, accompanying guardians, and previous recent physical examinations was recorded. Bivariate analyses and logistic regression were used to develop 2 sets of screening criteria to predict children at greatest risk of infection with: 1) GC and/or Ct (GC/Ct) and 2) GC alone. RESULTS: Fifty-eight children were identified with GC/Ct infections (37 with GC, 25 with Ct; 4 children were coinfected). The proposed algorithm for GC/Ct infections would have identified all children with these infections, while avoiding testing 56% of children without infection. Using genital cultures only, the proposed risk score for GC/Ct infections would have identified 100% of GC/Ct cases with 85% fewer cultures compared with testing all children with oral, rectal, and genital cultures. CONCLUSION: The use of a risk assessment algorithm for GC and Ct infections may reduce the cost and trauma of testing young children who are being evaluated for sexual abuse. sexual abuse, gonorrhea, Neisseria gonorrhoeae, chlamydia, Chlamydia trachomatis, sexually transmitted diseases, children.

Social context of sexual relationships among rural African Americans.

BACKGROUND: Reasons for the strikingly increased rates of HIV and other sexually transmitted infections (STIs) among African Americans in the rural Southeastern United States remain unclear. Investigators have devoted little attention to the potential influence of the social and economic context on sexual behaviors. GOAL: To examine the potential influence of these contextual factors on behaviors that promote the transmission of STIs. STUDY DESIGN: Focus group interviews in which African Americans from rural North Carolina discussed life in their communities and contextual factors affecting sexual behavior. RESULTS: Respondents reported pervasive economic and racial oppression, lack of community recreation, boredom, and resultant substance abuse. Many perceived a shortage of black men because of their higher mortality and incarceration rates compared with whites, and believed this male shortage to be partly responsible for the concurrent sexual partnerships that they perceived as widespread among unmarried persons. CONCLUSION: Contextual features including racism, discrimination, limited employment opportunity, and resultant economic and social inequity may promote sexual patterns that transmit STIs.

Violent behaviors in early adolescent minority youth: results from a "middle school youth risk behavior survey".

OBJECTIVES: To describe the prevalence and characteristics of violence and violence-related behaviors among six populations of U.S. minority adolescents in grades 6-8. METHODS: Six thousand four hundred non-White adolescents were recruited from six sites that were part of a collaborative project. Surveys were administered either during the school day or at community facilities. All students at each site were asked 10 questions about recent violence-related behaviors (including use of threats, fighting, weapon carrying, and weapon use). Prevalence of each violence-related behavior was reported within and across sites, and stratified by race/ethnicity, gender, age, and other characteristics expected to influence the behaviors. RESULTS: Sixty-six percent (66%) of the middle school students sampled reported being involved in some type of recent fighting and/or weapon-related behaviors. Sixty-one percent (61%) indicated some form of fighting behavior in the past 3 months (threatening to beat someone up, physical fighting, and/or being hurt in a fight). Thirty percent (30%) of participating youth reported one or more weapon-related behaviors (threatening to use a weapon, carrying a weapon, using a weapon, and/or being cut, stabbed or shot at). Reported gun carrying among males varied depending upon site, but was as high as 20%. Grade in school was positively associated with reported violent behaviors. Adolescents who reported living full-time with a parent or parent figure, and those who reported religious observance or beliefs, were less likely to report violence involvement. All violence-related behaviors were more common among male than female adolescents. CONCLUSIONS: Violence prevention efforts should begin in elementary school and continue throughout adolescence. Programs should be prepared to provide services or referrals to victims of violence, implement programs tailored toward females as well as males, and build partnerships with churches and other community organizations in which youth are involved.

Using north carolina medicare data to assess excess prostate cancer mortality among african americans.

PURPOSE: To investigate the basis for the higher prostate cancer mortality rate for African American (AA) men, which is twice the rate for White men.METHODS: 221 AA and 979 White men with a primary diagnosis code of prostate cancer ("patients") in the North Carolina Medicare Hospitalization claims from 1997 were compared with 1,326 AA and 5,874 White men of the same age with no cancer hospitalizations ("beneficiaries") selected from the NC Medicare Enrollment files. Mortality rates were calculated as the cumulative percent of deaths using the hospital discharge date as day 1. AA and White age distributions were similar.RESULTS: Cumulative mortality percentages at 6, 12, and 18 months were, respectively, 4.5, 7.7, 10.9 for AA patients; 2.8, 6.5, 9.2 for White patients; 2.3, 3.8, 7.4 for AA beneficiaries; and 1.8, 3.1, 6.1 for White beneficiaries.CONCLUSIONS: AA prostate cancer patients had higher overall mortality than did White prostate cancer patients during the first year, but by 12-months the White-Black survival advantage for prostate cancer patients was similar in magnitude to the White-Black survival advantage among the non-cancer Medicare beneficiaries. AAs' higher prostate cancer mortality may derive from higher short-term case fatality rates, which may reflect differences in treatment and access to quality medical care, co-morbidities, and tumor characteristics such as stage and grade at diagnosis, and in part from the survival disadvantage for AA in the general population.

Frequent re-infection in a community with hyperendemic gonorrhoea and chlamydia: appropriate clinical actions.

In a cohort of 1200 individuals experiencing an index infection of gonorrhoea or chlamydia in a rural North Carolina county, we studied the risk factors for re-infection within 16 months. The occurrence of asymptomatic re-infection was estimated from a sample of controls. Re-infection with either of the two STDs occurred in one-quarter (24%) of the study participants and 22% of the controls. The factors associated with re-infection were age less than 25 years, African-American race, male sex, and infection prior to the index infection. The behavioural and attitudinal variables tested were not associated with re-infection. The patterns of re-infection in this study do not provide guidance for targeting resources toward a subgroup of the clinic population. Structuring clinical services to make them more available and acceptable to all potential clinic clients can contribute to decreasing the duration of infections. Lowering the community prevalence in this way will then decrease the likelihood of re-infection rates.

The potential and limitations of data from population-based state cancer registries.

Cancer incidence varies markedly among states because of population heterogeneity regarding risk, genetic, and demographic factors. Population-based cancer registries are essential to monitoring cancer trends and control. The Centers for Disease Control and Prevention and the North American Association of Central Cancer Registries, through the National Program of Cancer Registries, are helping state registries generate more and better data nationwide. The National Program of Cancer Registries has supported the enhancement of 36 registries and the creation of 13 new registries in 45 states, 3 territories, and the District of Columbia, providing national standards for completeness, timeliness, and quality; financial support; and technical assistance. Users must be aware of diverse issues that influence collection and interpretation of cancer registry data, such as multiple cancer diagnoses, duplicate reports, reporting delays, misclassification of race/ethnicity, and pitfalls in estimations of cancer incidence rates. Attention to these issues and intense use of the available data for cancer surveillance will enable maximum societal benefit from the emerging network of population-based state cancer registries.

The African-American church: a potential forum for adolescent comprehensive sexuality education.

PURPOSE: To investigate the attitudes and beliefs of clergy from African-American churches towards sexuality education and the provision of sexuality education in their churches. METHODS: A pilot study was designed to survey a convenience sample of clergy leaders from African-American churches about their young adolescent members. The survey asked about priority health topics, prevalence of sexual and drug risk behavior and the clergy's desire for health education programs. The churches were located in a county (1990 population approximately 200,000, 40% African-American) in the southeastern United States. RESULTS: The respondents' highest priority issues were drugs, violence, HIV/AIDS, pregnancy and alcohol. Many (76%) had discussed one or more of these issues in church. All respondents wanted additional health seminars for their adolescents, though some clergy (30%) excluded some sexual topics (i.e., anal sex, bisexuality, homosexuality, masturbation, oral sex). Only 6% would make condoms available in their churches, but all would allow contraceptive education. CONCLUSIONS: Many African-American churches are open to including sexuality education among their health education offerings for young adolescents. The church should be considered as a potential forum for providing comprehensive sexuality education for African-American adolescents.

Trends in human immunodeficiency virus (HIV) counseling, testing, and antiretroviral treatment of HIV-infected women and perinatal transmission in North Carolina.

Since 1993, trends in perinatal human immunodeficiency virus (HIV) transmission have been monitored by use of chart review of patients identified at a central diagnostic laboratory. In the population studied, either pre- or postnatal antiretroviral therapy to the infant increased from 21% in 1993 to 95% in 1997. Concurrently, the number of HIV-infected infants declined from 25 in 1993 to 4 in 1997. The complete Pediatric AIDS Clinical Trials Group Protocol 076 regimen was the most effective in reducing transmission (3.1%). Twenty-two of 35 infants who became infected in 1995-1997 had mothers who did not receive antiretroviral therapy, although counseling practices improved with time. In 1995, 87% of the mothers of HIV-seropositive infants were counseled, whereas in 1997, 96% were counseled (P<.005). None of 59 infants tested had high-level phenotypic zidovudine resistance, although 5 (8.8%) of 57 infants had virus isolates with at least one mutation in the reverse transcriptase gene associated with reduced phenotypic susceptibility to zidovudine.

The epidemiology of adolescent homicide in North Carolina from 1990 to 1995.

BACKGROUND: Rates of homicides by adolescents under age 18 years tripled from 1984 to 1994. Most studies report data on urban adolescents and young adults as a single age group (age 15-24 years), but homicide characteristics among adolescents, especially those younger than 15 years, may differ from those of young adults. OBJECTIVE: To describe the homicide characteristics among adolescents age 11 to 18 years in North Carolina from 1990 to 1995. METHODS: A retrospective, descriptive analysis of adolescent homicides using the medical examiner database. Police interviews provided additional information for cases from 1993 to 1995. RESULTS: There were 419 victims from 1990 to 1995 (average annual rate: 9.7 per 100000 adolescents; 9.9 in urban counties, 7.1 in rural). Victims were mostly ages 15 to 18 years (85%), male (79%), and black (76%); 48% lagged behind in school, and, by police report, 40% had a criminal record. Only 23% of the identified perpetrators were strangers. Firearms (59% were handguns) were used in 83% of homicides. Proportionally more younger adolescents (age 11-14 years) were killed by means other than firearms than 15- to 18-year-olds (chi2 = 24.2, P = .007). Drug-related motives (23%) were most common, followed by non-drug-related altercations (20%) and retaliations (17%). CONCLUSIONS: Proportionally more North Carolina adolescents than urban young adults (ages 15-24 years) were killed by firearms (83% vs. 75%). Proportionally fewer adolescents were killed by police, strangers, or intimate partners. Interventions should include reducing access to firearms and drugs, and helping adolescents develop nonviolent strategies to resolve disputes. Efforts should be focused on adolescents who lag behind in school and have criminal records.

Influence of cycle variability and coital frequency on the risk of pregnancy.

Researchers have cautioned against generalizing results from contraceptive trials because these studies rely on self-selected participants meeting strict selection criteria who may differ from typical users. Using information collected on daily diaries, we reanalyzed data from the recently completed Reality female condom clinical trial to evaluate factors that influence the probability of pregnancy. Noncompliant women, women with less variable menstrual cycles (17-43 days), and women engaging in intercourse frequently (> or = 11 acts per month) were more likely to conceive during this 6-month trial. The adjusted hazard ratios and 95% confidence intervals for these three covariates were 6.1 (2.0-18.7), 7.2 (1.0-54.3), and 2.0 (0.7-5.3), respectively. The strict selection criteria used in this study failed to recruit a homogeneous cohort with respect to factors that influence the risk of pregnancy. The overall pregnancy rate does not pertain to individual study participants, but rather represent average effects for a population with the particular mix of characteristics found in this study. In particular, we not only confirm the well known importance of compliance and the obvious role of frequency of intercourse, but also demonstrate that women with cycles outside the range of 17-43 days appear to be at a much lower risk of pregnancy.

PIP: This article highlights the influence of cycle variability and coital frequency on the risk of pregnancy using data from the completed female condom trial. Daily diaries of women, which include coital frequency and menstrual cycle, were reanalyzed to evaluate the factors that affect the probability of pregnancy. A higher probability of pregnancy was reported among noncompliant women, women with less variable menstrual cycles, and women engaging in frequent intercourse during a 6-month trial. The adjusted hazard ratios and 95% confidence intervals for these three covariates were 6.1 and 2.0-18.7 (noncompliant women), 7.2 and 1.0-54.3 (less variable menstrual cycles), and 2.0 and 0.7-5.3 (frequent intercourse), respectively. The overall pregnancy rate obtained in this study represent average effects for a population with the particular mix of characteristics. This study appears to demonstrate a much lower risk of pregnancy for women with menstrual cycles outside the range of 17-43 days.

Factors contributing to the poorer survival of black Americans diagnosed with oral cancer (United States).

OBJECTIVE: The purpose of this study was to identify factors that contribute to the poorer survival of blacks in the United States diagnosed with oral cancer. METHODS: Data for 6,338 whites and 1,165 blacks diagnosed from 1988 to 1993 with squamous cell carcinoma of the oral cavity and pharynx were obtained from the National Cancer Institute's Surveillance, Epidemiology and End Results (SEER) program 1973-1993 Public-Use Database. The covariables were sex, age, geographic area, marital status, socioeconomic status (five census-tract measures), stage, anatomic site, grade, lymph node involvement, tumor size, and treatment. Hazard ratios were estimated with Cox regression. RESULTS: Adjusted for age and geographic area, the hazard of death from oral cancer was 1.7 (95% confidence interval: 1.5-1.9) times greater among blacks than whites. The addition of the socioeconomic status (SES) variables to the model reduced the hazard ratio for race to 1.3 (1.0-1.7). Further adjustment by stage and treatment reduced the hazard ratio for race to 1.1 (0.9-1.4). In a model containing all covariables (except lymph node involvement and tumor size), the hazard ratio for race remained 1.1 (0.9-1.4). Analyses with the outcome death from any cause gave similar results. CONCLUSIONS: Lower SES, more advanced stage, and differences in treatment accounted for 86% of the excess hazard of death from oral cancer among blacks.

Recruitment of minority students to U.S. epidemiology degree programs. The American College of Epidemiology Committee on Minority Affairs.

PURPOSE: African-, Hispanic-, and Native Americans are underrepresented in the field of epidemiology including degree programs. As part of the assessment component of its mandate, the American College of Epidemiology Committee on Minority Affairs conducted a survey of minority recruitment activities of U.S. epidemiology degree programs. METHODS: The survey, containing questions related to marketing activities, institution infrastructure, financial support, academic offerings, and receptive/supportive environment, was mailed to all programs identified in Episource as offering epidemiology degrees. Separate responses were requested concerning activities at the department and school levels. RESULTS: Fifty-two completed questionnaires were received (response rate of 79%). All but two institutions had at least one activity conducted by either the department or the school. However, all activities were more common at the school- than at the department-level. Indeed, some activities [a written minority student recruitment plan (6% of departments and 52% of schools), personnel with minority recruitment responsibilities (4% of departments and 73% of schools)] were almost exclusively school-sponsored. Although marketing-type activities were the most common minority recruitment tool used by departments, only 21% made visits to minority schools, 17% visited other colleges specifically to recruit minorities, and 12% produced materials targeted to ethnic/racial minorities. Six percent of the departments and 19% of the schools offered financial support (grants, fellowships, scholarships) to almost all underrepresented minority students. CONCLUSIONS: Even though individual epidemiology degree programs may not see a need for general recruitment activities in order to maintain the size of their applicant pool, minority-specific recruitment activities should be undertaken to enhance and diversify that pool. We recommend that epidemiology departments develop, adopt, and implement comprehensive written plans for the recruitment of underrepresented minority students into their programs.

Perinatal HIV infection and the effect of zidovudine therapy on transmission in rural and urban counties.

OBJECTIVES: To assess health care providers' identification of human immunodeficiency virus (HIV)-exposed infants, to ascertain the prevalence of transplacental or oral zidovudine treatment among infants exposed to HIV, and to estimate the impact of zidovudine use on perinatal transmission in rural and urban North Carolina. DESIGN: Survey of North Carolina newborns tested for HIV infection in 1993 and 1994 compared with the number of anonymous HIV-positive childbearing women. SETTING: North Carolina hospitals, public health clinics, and private physicians' offices. MAIN OUTCOME MEASURES: Rates of identification of HIV-exposed infants and of perinatal HIV-1 transmission, determined by HIV culture and polymerase chain reaction testing in the infants. RESULTS: The proportion of HIV-exposed children in North Carolina who were identified and tested increased from 60% in 1993 to 82% for all of 1994, and to more than 90% for the last quarter of 1994. The HIV-exposed infants born in rural counties were more likely to be recognized than those born in urban counties (P<.001). In 1994, most infants were evaluated relatively early in life: 39% by 1 week of age, 63% by 6 weeks, and 76% by 3 months. Among infants with recognized HIV exposure, transmission decreased significantly between 1993 and 1994, from 21% to 8.5%, respectively (P=.009). After the announcement of the results of the AIDS Clinical Trials Group Protocol 076, zidovudine was given to 75% of HIV-positive women who delivered infants in North Carolina. Only 5.7% of infants who received any zidovudine became infected, compared with 18.9% of infants who received no zidovudine (P=.007). CONCLUSIONS: Health care providers in North Carolina are identifying most of the state's HIV-seropositive pregnant women, treating them with zidovudine, and testing their infants soon after birth for HIV infection. The use of zidovudine in pregnant women and their infants has reduced perinatal HIV transmission in the state.