ORIGINAL

ARTICLES

A Practical Risk Measure for

Identification of Adults With Oncogenic
Oral Human Papillomavirus: Potential
Use in Primary Care
Arch G. Mainous III, PhD; Young-Rock Hong, MPH; Sandhya Yadav, MHA; Zhigang
Xie, MPA; Rebecca J. Tanner, MA

BACKGROUND AND OBJECTIVES: Oral human papillomavirus (HPV) infection is aged 18 to 64 years did not see a
the main cause of oropharyngeal cancer. However, there is no assessment tool for dentist in 2016.7 Consequently, be-
early detection and prevention of oropharyngeal cancer in practice. The purpose of cause of the relationship between oral
the study was to develop and validate a risk assessment tool to pre-dict the health and systemic disease,8-10 primary
presence of HPV associated with oropharyngeal cancer. care providers have become more
aware of oral health screen-ing and
METHODS: Using data from the National Health and Nutrition
examinations. The Smiles for Life
Examination Survey 2011-2014, 6,978 US adults aged 18 to 59 years who
program is an example of a very
were tested for oral HPV infection were included for this study. We carried
out an analysis to test and validate risk predictive models for oral HPV successful oral health cur-riculum for
infection. Presence of one of the 20 HPV subtypes associated with primary care clinicians
oropharyngeal cancer was used for the outcome. (https://smilesforlifeoralhealth.org). It
includes oral health risk assess-ment
RESULTS: Of 6,978 participants aged 18-59, 303 (4.3%; 6.6 million) were tools. However, it does not in-clude
found to have oncogenic HPV subtypes. Our final model included sex, income- strategies to identify patients at high
to-poverty ratio, current smoking, and the lifetime number of oral sex partners. risk for oral HPV.
The discriminatory power of the oral HPV risk score to predict the presence of Currently, screening for oncogenic
oncogenic HPV was good (C-statistic=0.73). The risk score performed compa- oral HPV would be challenging for
rably in the validation population (C-statistic=0.72). The comparison between
several reasons. First, population-wide
observed and estimated proportions of population with oncogenic oral HPV
screening of all asymptomat-ic adults
demonstrated excellent calibration.
would be very inefficient, realizing that
CONCLUSIONS: We developed and validated the oral HPV risk score that pre- only a small propor-tion of the adult
dicts the risk of oral HPV requiring only self-reported data and no laboratory testing. population is in-fected. Second, the
The Oral HPV risk score has the potential to provide clinicians with a no-cost, easy cost of the test is not inconsequential at
way to screen for patients at greater risk for oncogenic HPV infection. approxi-mately $90
(https://www.lifelabs.com/ how-to-
(Fam Med. 2020;52(1):31-7.) order-hpv/), suggesting that targeted
doi: 10.22454/FamMed.2020.382138 strategies for detection or increased
surveillance for oropha-
Human papillomavirus (HPV) has been
identified as a key cause in the ryngeal cancer would be preferred.
development of common oncogenic oral HPV (sub-
Further, the US Preventive Servic-es
oropharyngeal cancer.1 In fact, HPV is type 16) is present in 3.5% of US
Task Force recommendation on
suggested to be the underlying cause of adults.4 Moreover, although many of
screening for oral cancer states that
70% of oropharyngeal can-cers in the those infections clear, a substan-tial
United States.2 Oral HPV is not number of individuals have per-sistent
uncommon.3 However, not all subtypes infections.5 From the Department of Health Services
Many individuals do not have op- Research, Management, and Policy (Dr Mainous,
of HPV are linked to oro-pharyngeal Mr Hong, and Mss Yadav, Xie, and Tanner); and
cancer. One of the most timal access to dental services,6 the Department of Community Health and Family
National Center for Health Statistics Medicine, University of Florida, Gainesville, FL
indicated that 35.6% of US adults (Dr Mainous).

FAMILY MEDICINE VOL. 52, NO. 1 • JANUARY 2020 31

 ORIGINAL ARTICLES

there is insufficient evidence to rec- HPV vaccine would be at very low the past 30 days.14 Sexual behavior
ommend screening in asymptomatic risk, and so a risk score would not be information included lifetime num-ber
adults.11 With no specific guidance for useful for them. Based on these of oral sex partner regardless of gender
screening for early detection and exclusion criteria, our final analyt-ic orientation and history of any sexually
prevention of oropharyngeal can-cer, it sample included 6,978 US adults aged transmitted infection (eg, genital warts,
would be useful to identify the 18-59 years, representing pop-ulation herpes, gonorrhea, and chlamydia).
population most at risk for car-riage of estimates for 153,886,760 Americans. Oral sex partners was classified into
oncogenic HPV because of its outsized two groups (0-7 part-ners and at least 8
role in the development of partners) with the high group
oropharyngeal cancers. Thus, the Measures corresponding to the 75th percentile
purpose of this study was to develop a To be adopted into clinical practice, in among the sample of adults.
risk score for oncogenic oral HPV addition to being valid, a predic-tion
based on patient self-reported char- tool must be intuitive and con-
acteristics. We will develop and vali- ceptually simple.12 The conceptual Outcome Variable
date the risk score using nationally basis for the oral HPV risk score is to The outcome variable is the presence
representative data for the United build on information that is read-ily of one of the 20 HPV subtypes as-
States. This goal is to create a prac- obtainable from the patient by self- sociated with oropharyngeal cancer.
tical and easily-implemented risk report. We only examined pre-dictors We added eight more subtypes sug-
stratification tool that could identi-fy that could be self-report-ed or gested by literature in addition to the
individuals who may benefit from measured without obtaining a group 1 carcinogens, resulting in the
more frequent oral exams for preven- laboratory sample, because if a inclusion of HPV subtypes 6, 11, 16,
tion of oropharyngeal cancer. laboratory sample was required for 18, 31, 32, 33, 35, 39, 45, 51, 52, 53,
computation, it would diminish the 56, 57, 58, 59, 73, 81, and 82.15-17 Oral
Methods ease and value of the risk score in a rinse specimens were processed,
Data and Sample clinical context. stored, and shipped to a central labo-
The development and validation of the ratory. Detailed specimen collection
oral HPV risk score was under-taken Candidate Variables and processing instructions are dis-
using data from adults aged 18-59 We selected the candidate variables cussed in the NHANES Laborato-
years, from the population-based from risk factors for oral HPV identi- ry/Medical Technologists Procedures
National Health and Nutrition fied from the literature. 3,13 We identi- Manual.18 Purified DNA was ana-lyzed
Examination Survey (NHANES), fied the variables from the literature for 37 types of HPV by means of a
2011-2014. The NHANES is a mul- that met our criteria of being able to be multiplex polymerase-chain-re-action
tistage probability sample of the assessed via patient self-report; we (PCR) assay. For the oral HPV risk
civilian, noninstitutionalized US recoded the variables to simplify and score we classed as our outcome
population. The NHANES includes result in meaningful categories prior to variable presence of one or more of the
physical exams, laboratory tests, and analysis. Sociodemographic 20 HPV subtypes.
interviews with participating indi- characteristics included age (18-39
viduals. The Institutional Review years and 40-59 years); sex (female Data Analyses
Board of the University of Florida and male); race/ethnicity (non-His- Model and Risk Score Develop-ment.
approved this study; the study was panic white, non-Hispanic black, His- All analyses were conduct-ed using
conducted in 2018. panic, and other); education (high SAS 9.4 (SAS Institute) so that we
Our initial unweighted sample school diploma or less and at least could make population es-timates for
included 9,853 participants in the some college); marital status (mar-ried the US noninstitution-alized adult
NHANES mobile examination cen-ter and not married), income-to-poverty population. We took the dataset based
(MEC). Participants more than 59 ratio (IPR; <1.0 [in poverty] and ≥ 1.0 on the 2011-2014 NHANES and
years old (n=1,786) were ex-cluded [not in poverty]). Concern-ing randomly split it into two equal
from our risk score develop-ment substance use, we included and used databases. One database was used for
because some questions about sexual dichotomized measures of cur-rent model and score devel-opment to
behavior were not asked of these smoking, marijuana use, and binge determine effectiveness of the new risk
persons. Individuals who in-dicated drinking. Current marijuana use was score classification in identifying
that they had received the HPV vaccine defined if a participant re-ported using individuals with carcino-genic oral
were not included in the risk score marijuana or hash-ish during the past HPV. The holdout sample was used for
development analysis (n=1,089). 30 days. Binge drinking was defined as additional analysis to determine and
Although the current HPV vaccines do consump-tion of at least five five validate whether, in a new sample,
not cover all of the carci-nogenic HPV drinks for male or at least four for subjects can be correctly classified
subtypes, we felt that individuals who female in 2 hours at least one occasion with respect to oral HPV.
had received the during

32 JANUARY 2020 • VOL. 52, NO. 1 FAMILY MEDICINE

 ORIGINAL ARTICLES

We used multivariable logistic indicates perfect calibration). The (Figure 2-B). There was no evidence of
regression. Beginning with the full sensitivity, specificity, positive pre- risk overestimates with calibra-tion
unrestricted model including all dictive value (PPV), negative predic- slopes 1.06 and 1.03, respec-tively,
candidate variables, we used man-ual tive value (NPV), positive likelihood suggesting that the developed score
backwards elimination to deter-mine ratio (+LR), and negative likelihood was well calibrated. The perfor-mance
the optimal set of candidate variables, ratio (-LR) were reported for various of the risk score at different cut-off
which would best predict the presence cutoff points of the final risk score. points is present in Appen-dix Table 1
of carcinogenic oral HPV. We treated (https://journals.stfm.org/
missing values for the candidate Results media/2805/mainous-appendixtable-
variables (rang-ing from 2.9% for Among 6,978 participants aged 18-59 fm52-1-2020.pdf). A score of at least 5
smoking to 13.8% for the lifetime who were tested for oral HPV in- (85.9% vs 47.3%) or at least 7 (78.9%
number of oral sex partners) by listwise fection, 303 (4.3%; 6.6 million) were vs 56.8%) had a higher sen-sitivity,
deletion in the models. We compared with the presence of HPV subtypes while a score of 8 (86.4% vs 57.0%) or
each step of variable exclusions (nested associated with oropharyngeal can-cer. greater cut-off points had a higher
models) with the full model and Table 1 shows the baseline char- specificity than sensitivity. Cut-off
preceding model using the negative of acteristics of the participants with and points of at least 8 or 9 yield-ed
twice the log likelihood (-2 LL) and without HPV subtypes associat-ed with comparable performance with 20% to
Aikaike Information Criteria (AIC). oropharyngeal cancer. Partic-ipants 30% increased probabilities of oral
We re-ported odds ratios (ORs) and with HPV infection associated with HPV infection (ie, +LR of 2 in-crease
95% confidence intervals (CIs) for the oropharyngeal cancer were more likely probability 15%).20 For clinical use,
fi-nal multivariable model. to be male, non-Hispanic black, and different cutoff points may be chosen
had a poor health-related behavior depending on screening pur-poses:
profile (eg, current smoker, marijuana high sensitivity for minimiz-ing false
Converting the Model Results to use, binge drinking, and greater negatives vs high specificity for
Integer Risk Points. The mea-sure is number of oral sex partners). The oral minimizing false positives.
designed to be conceptual-ly easy to HPV risk score was derived from the
grasp by clinicians and used in clinical first 3,489 participants and validated Discussion
settings as well as by patients on their from the second half. No significant We developed and validated a simple
own. For this reason, we rounded the differences across base-lines integer-point risk score using a na-
resulting sta-tistically significant odds characteristics were observed between tionally representative cohort. This is
ratios to whole number risk points, the development and vali-dation the first study to create a risk score to
using the method of Charlson in cohorts. identify those most at risk for on-
developing the Charlson Comorbidity After the backward elimination of cogenic oral HPV, which would guide
Index.19 Using this strategy, significant the nonsignificant candidate vari-ables, the screening for early detection and
odds ratios between 1 and 1.19 scored the final best fit model includ-ed sex, prevention of oropharyngeal cancer
0 points; ratios from 1.2 to 1.49 were IPR, current smoking, and the lifetime caused by HPV. The risk score pre-
rounded to 1 point; ratios from 1.5 to number of oral sex part-ners (Table 2). dicts the risk of oral HPV requiring
2.49 were rounded to 2 points, and so Being male scored 5 (OR, 5.18, 95% only self-reported data and no lab-
forth. The regressions were com-puted CI, 2.96-9.07), in poverty scored 2 oratory testing. Because of the use of
with the risk scores based on integer (OR, 1.93, 95% CI, 1.06-3.52), current only self-reported data and no re-
risk points. We computed C-statistics smoking scored 2 (OR, 2.31, 95% CI, quired diagnostics, this risk score has
and evaluated to assess the 1.46-3.67), and having more than 7 potential utility in public health, health
discrimination and classification util- lifetime oral sex partners scored 3 (OR, systems and self-evaluation settings
ity of the model. 3.06, 95% CI, 1.73-5.40). The area with little to no financial cost and very
under the curve (AUC) for this model little time required.
was 0.730 (C-statistic), indicating The value of this easily computed
Risk Score Validation. We com-puted adequate discrimination (Figure 1-A). risk score is its widespread applica-
the risk scores in the holdout sample The per-formance of oral HPV risk bility that allows for risk stratifica-tion
used for validation, and we compared score in the validation cohort was of individuals and appropriate
the resulting C statistics between the similar giv-en C-statistic of 0.721. surveillance for early detection of
development sample and the validation Figure 2 show the calibration plots for oropharyngeal cancer and target-ed
sample. Dis-crimination and calibration observed proportion of participants counseling efforts for behavioral
in the validation sample was then with oral HPV infection and predicted modification of high-risk individu-als.
tested by computing the C‐statistic and risks across the risk score points It can be particularly useful for guiding
calibration slope and intercept (ie, a (Figure 2-A) and predicted risks people who do not go to the dentist for
slope of 1 and an intercept of 0 between the development and routine oral examinations. The risk
validation cohorts score can be administered

FAMILY MEDICINE VOL. 52, NO. 1 • JANUARY 2020 33

 ORIGINAL ARTICLES

Table 1: Characteristics of US Population With and Without Carcinogenic

Oral HPV (n=6,978, Weight N=153 Million Americans)
No OR-HPV OR-HPV P Development Validation
Variables
(Carcinogenic) (Carcinogenic) Value Cohort Cohort
Sample n=6,675 Sample n=303 3,489 3,489
Weighted Weighted
75,505,927 78,380,833
N=147,263,235 N=6,623,526
Total
No. (Row %) b No. (Row %) b No. (Col %) b No. (Col %) b
No. a
Age .145
18-39 yrs 3,,475 3,331 (96.1) 144 (3.9) 1,733 (47.5) 1,742 (46.3)
40-59 yrs 3,503 3,344 (95.3) 159 (4.7) 1,756 (52.5) 1,747 (53.7)
Sex <.001
Male 3,507 3,259 (92.8) 248 (7.2) 1,771 (51.7) 1,753 (50.2)
Female 3,471 3,416 (98.6) 55 (1.4) 1,718 (48.3) 1,736 (49.8)
Race/Ethnicity .003
NH-white 2,577 2,446 (95.4) 131 (4.6) 1,273 (61.8) 1,304 (63.0)
NH-black 1,616 1,533 (94.9) 83 (5.1) 858 (13.2) 758 (11.2)
Hispanic 1,559 1,502 (96.5) 57 (3.5) 757 (16.7) 802 (17.0)
Other 1,226 1,194 (97.4) 32 (2.6) 601 (8.3) 625 (8.8)
Education .006
HS or less 3,126 2,959 (94.7) 167 (5.3) 1,583 (40.2) 1,543 (36.9)
Some college or higher 3,850 3,714 (96.3) 136 (3.7) 1,905 (59.8) 1,945 (63.1)
Marital Status .002
Not married 3,264 3,090 (94.3) 174 (5.7) 1,644 (44.9) 1,620 (46.1)
Married 3,343 3,224 (96.7) 119 (3.3) 1,674 (55.1) 1,669 (53.9)
Poverty Level .810
Yes (IPR<1.0) 1,640 1,562 (95.8) 78 (4.2) 812 (18.0) 828 (18.0)
No (IPR≥1.0) 4,750 4,549 (95.6) 201 (4.4) 2,368 (82.0) 2,382 (82.0)
Binge Drinking (4/5 drinks
.013
or more once a month)
No 5,008 4,814 (96.1) 194 (3.9) 2,489 (79.6) 2,519 (78.9)
Yes 1,132 1,053 (93.9) 79 (6.1) 568 (20.4) 564 (21.1)
Current Marijuana Use (ever
.005
used in the past month)
No 5,140 4,930 (96.0) 210 (4.0) 2,550 (86.4) 2,590 (85.6)
Yes 901 838 (93.1) 63 (6.9) 449 (13.6) 452 (14.4)
Current Smoking <.001
No 5,149 4,976 (96.9) 173 (3.1) 2,573 (77.3) 2,576 (76.3)
Yes 1,628 1,506 (91.6) 122 (8.4) 823 (22.7) 805 (23.7)
# of Oral Sex Partners in Lifetime <.001
0-7 5,093 4,935 (97.3) 158 (2.7) 2,506 (81.8) 2,587 (81.4)
8 or More 910 801 (88.4) 109 (11.6) 474 (18.2) 436 (18.6)
Had STDs/STIs in the Past 12 Months .782
No 5,237 4,996 (95.5) 241 (4.5) 2,623 (89.9) 2,614 (89.7)
Yes 490 467 (95.8) 23 (4.2) 233 (10.1) 257 (10.3)
a Total sample size is not equal due to missing values.
b Percentages and 95% CI are weighted to be nationally representative.

34 JANUARY 2020 • VOL. 52, NO. 1 FAMILY MEDICINE

 ORIGINAL ARTICLES

Table 2: Multiple Logistic Regression Results and Risk Score Assignment

Variables Odds Ratio 95% CI P Value Score Points
Sex
Female 1.00 0
Male 5.18 2.96 to 9.07 <.001 5
Poverty Level
No (IPR≥1.0) 1.00 0
Yes (IPR<1.0) 1.93 1.06 to 3.52 .032 2
Current Smoking
No 1.00 0
Yes 2.31 1.46 to 3.67 .001 2
No. of Oral Sex
Partners in Lifetime
0-7 1.00 0
8 or More 3.06 1.73 to 5.40 <.001 3
Total Points
12

Abbreviation: IPR, income to poverty ratio.

Figure 1: Receiver Operating Characteristic (ROC) Curves for Oral-HPV Risk Score
and Comparison Between Development (A) and Validation (B) Cohorts

FAMILY MEDICINE VOL. 52, NO. 1 • JANUARY 2020 35

 ORIGINAL ARTICLES

Figure 2: Comparisons Between Development and Validation Cohorts (External Calibration; A)

and Distributions of Observed and Predicted Risk for Oral HPV (Internal Calibration; B)

in the primary care physician’s wait- detection and prevention, a score of 7 examines the current risk and does not
ing room and other outpatient set-tings, or higher (78.9% sensitivity and 56.8% predict future risk since it has been
or individuals can calculate their risk at specificity) may be chosen un-der the developed from cross-section-al data.
home. After self-eval-uation, if an current context of the United States The C-statistics in both de-velopment
individual finds him-self in the high- when there are no national guidelines and validation model is around 0.73,
risk category, he/she can be mindful of for screening of oncogen-ic oral HPV. which is not the best; however, it is
any mouth lesion that may appear and Nevertheless, when ap-plied to a consistent with other similar cross-
that takes lon-ger to heal, and schedule specific setting, it should be noted that sectional risk scores developed to
an appoint-ment with the dentist for a the chosen threshold needs to be guide screening and surveillance. 23-25
timely examination. Given the decided in consideration of other Although self-re-ported characteristics
increasing emphasis on nondental factors such as the preva-lence of increase the applicability of this tool,
health care providers in oral health care disease in the target popu-lation, the self-report itself suffers from recall
ser-vices,8-10,21 family physicians and resources available, and the screening bias and so-cial desirability,
primary care providers should be en- requirements. particularly around the sensitive topic
couraged to play a role in oral cancer Since the risk score was developed of sexual behav-ior. Further, although
screening during their routine physi-cal using a population-based nationally the poverty-to-income ratio is a better
exams, including quick visual in- representative sample in the United indicator of economic status than
spection (eg, oral cavity, tongue, and States, its validity to be applicable household in-come because it includes
oropharynx for suspicious legions), nationwide increases. Also, we only both family size and the federal
oral health counseling, and further used factors that were self-reported or poverty thresh-old, it would be hard for
referrals based on the results of the measured without any laboratory test patients to compute immediately.
screening. It is also critical for them to and converted each risk factor to Providing some help in the clinic to the
promote HPV vaccination uptake simple integer scores, which makes its patient to tell them if they are above or
among age-eligible adolescents and applicability simple, cost-effec-tive, be-low the poverty line may be neces-
young adults through patient educa- and easy in both clinical and self- sary. The predictability of this risk
tion (including parents or guardians) to evaluation settings. Additional-ly, this score might be fluctuated among
raise awareness of HPV-related health tool can also be administered to screen different oropharyngeal cancers due to
consequences.22 the population with lower utilization of the differential gradient of the strength
dental care to improve referrals and of association between HPV and that
Considering the untreatable na-ture routine follow-ups for surveillance. specific cancer. Lastly, the current
of HPV infection, unfavorable application of the tool is limited only to
prognosis of oropharyngeal cancer, and Despite the many strengths of this the population of the United States
the low cost and high poten-tial study, some limitations need to be with ages between 18
benefits of screening for early acknowledged. This risk score only

36 JANUARY 2020 • VOL. 52, NO. 1 FAMILY MEDICINE

 ORIGINAL ARTICLES

and 59. For example, considering the 6. Sams LD, Rozier RG, Quinonez RB. Training 17. Shigeishi H, Sugiyama M. Risk Factors for Oral
median age of 63 for oropharyngeal Requirements and Curriculum Content for Human Papillomavirus Infection in Healthy
Primary Care Providers Delivering Preventive Individuals: A Systematic Review and Meta-
cancer, further external validation to Oral Health Services to Children Enrolled in Analysis. J Clin Med Res. 2016;8(10):721-729.
apply to those aged 60 or older is Medicaid. Fam Med. 2016;48(7):556-560.
needed. 7. Center for Disease Control and Prevention. Oral 18. Centers for Disease Control and Prevention.
This is the first risk score that al- and Dental Health. 2016. https://www.cdc. NHANES Laboratory/Medical Technologists
gov/nchs/fastats/dental.htm. Accessed October Procedures Manual. 2013. https://wwwn.
lows for risk stratification of adults for 25, 2019. cdc.gov/nchs/data/nhanes/2013-2014/manu-
oncogenic oral HPV and can be applied 8. Institute of Medicine and National Research als/2013_MEC_Laboratory_Procedures_Man-
in routine clinical practice for early Council. Improving Access to Oral Health Care ual.pdf. Accessed January 7, 2018.

detection of oropharyn-geal cancer and for Vulnerable and Underserved Populations. 19. Charlson ME, Pompei P, Ales KL, MacKenzie
Washington, DC: National Academies Press; CR. A new method of classifying prognostic co-
targeted behavior-al modification. 2011. morbidity in longitudinal studies: development
Further research is warranted to 9. Health Resources and Services Administration. and validation. J Chronic Dis. 1987;40(5):373-
investigate the timing and efficacy of Integration of Oral Health and Primary Care 383.
Practice. U.S. Department of Health and Hu-man 20. McGee S. Simplifying likelihood ratios. J Gen
screening in different settings that
Services, Health Resources and Services Intern Med. 2002;17(8):646-649.
contribute to the valid-ity of this risk Administration, February 2014. https://www. 21. Cohen LA. Expanding the physician’s role in
score. hrsa.gov/sites/default/files/hrsa/oralhealth/in- addressing the oral health of adults. Am J Public
tegrationoforalhealth.pdf. Accessed October 25, Health. 2013;103(3):408-412.
CORRESPONDING AUTHOR: Address corre- 2019.
spondence to Dr Arch G. Mainous III, Depart- 22. American Academy of Pediatric Dentistry Policy
10. Association of American Medical Colleges. on human papilloma virus vaccina-tions.
ment of Health Services Research, Manage-ment Report IX Contemporary Issues in Medicine:
and Policy, University of Florida Health Sciences Chicago: AAPD; 2017. http://www.aapd.
Oral Health Education for Medical and Dental org/media/Policies_Guidelines/P_HPV_Vaccina-
Center, PO Box 100195, Gainesville, FL. 352- Students. Washington DC: AAMC; June 2008.
273-6073. [email protected]. tions.pdf. Accessed October 25, 2019.
11. U.S. Preventive Services Task Force. Clini-cal 23. Barengo NC, Tamayo DC, Tono T, Tuomilehto
Summary. Oral Cancer: Screening. https:// J. A Colombian diabetes risk score for detecting
References www.uspreventiveservicestaskforce.org/Page/ undiagnosed diabetes and impaired glucose
Document/ClinicalSummaryFinal/oral-cancer- regulation. Prim Care Diabetes. 2017;11(1):86-
1. Candotto V, Lauritano D, Nardone M, et al. HPV
screening1. Accessed December 14, 2018. 93.
infection in the oral cavity: epidemiol-ogy,
clinical manifestations and relationship with oral 12. Liao L, Mark DB. Clinical prediction models: 24. Haroon S, Adab P, Riley RD, Fitzmaurice D,
cancer. Oral Implantol (Rome). 2017;10(3):209- are we building better mousetraps? J Am Coll Jordan RE. Predicting risk of undiag-nosed
220. Cardiol. 2003;42(5):851-853. COPD: development and validation of the
2. Chaturvedi AK, Engels EA, Pfeiffer RM, et al. 13. D’Souza G, Wentz A, Kluz N, et al. Sex Differ- TargetCOPD score. Eur Respir J.
Human papillomavirus and rising oropharyn-geal ences in Risk Factors and Natural History of Oral 2017;49(6):1602191.
cancer incidence in the United States. J Clin Human Papillomavirus Infection. J Infect Dis. 25. Carrillo-Larco RM, Miranda JJ, Gilman RH, et
Oncol. 2011;29(32):4294-4301. 2016;213(12):1893-1896. al; CRONICAS Cohort Study Group. Risk score
3. Gillison ML, Broutian T, Pickard RK, et al. 14. National Institute on Alcohol Abuse and Alco- for first-screening of prevalent undi-agnosed
Prevalence of oral HPV infection in the United holism. Drinking Levels Defined: Binge Drink- chronic kidney disease in Peru: the CRONICAS-
States, 2009-2010. JAMA. 2012;307(7):693-703. ing. https://www.niaaa.nih.gov/alcohol-health/ CKD risk score. BMC Nephrol. 2017;18(1):343.
overview-alcohol-consumption/moderate-binge-
4. D’Souza G, McNeel TS, Fakhry C. Understand-
drinking. Published 2018. Accessed January 1,
ing personal risk of oropharyngeal cancer: risk-
2019.
groups for oncogenic oral HPV infec-tion and
oropharyngeal cancer. Ann Oncol. 15. Bouvard V, Baan R, Straif K, et al; WHO In-
2017;28(12):3065-3069. ternational Agency for Research on Cancer
Monograph Working Group. A review of hu-
5. Pierce Campbell CM, Kreimer AR, Lin HY, et
al. Long-term persistence of oral human man carcinogens—Part B: biological agents.
Lancet Oncol. 2009;10(4):321-322.
papillomavirus type 16: the HPV Infection in
Men (HIM) study. Cancer Prev Res (Phila). 16. Kim SM. Human papilloma virus in oral cancer.
2015;8(3):190-196. J Korean Assoc Oral Maxillofac Surg.
2016;42(6):327-336.

FAMILY MEDICINE VOL. 52, NO. 1 • JANUARY 2020 37