Journal of Public Health Dentistry . ISSN 0022-4006
Physicians’ attitudes toward medical screening in a dental setting Barbara L. Greenberg, MSc, PhD1; Pauline A. Thomas, MD2; Michael Glick, DMD3; Mel L. Kantor, DDS, MPH, PhD4 1 2 3 4
School of Health Sciences and Practice New York Medical College, Valhalla, NY, USA New Jersey Medical School, Rutgers University, Newark, NJ, USA School of Dental Medicine, University at Buffalo, Buffalo, NY, USA College of Dentistry and College of Public Health, University of Kentucky, Lexington, KY, USA
Keywords attitudes; screening; diabetes; heart disease. Correspondence Dr. Barbara L. Greenberg, School of Health Sciences and Practice New York Medical College, 40 Sunshine Cottage Road, Valhalla, NY 10595. Tel.: 914-594-3635; Fax: 914-594-4853; e-mail: barbara_greenberg @nymc.edu. Barbara L. Greenberg is with the School of Health Sciences and Practice New York Medical College. Pauline A. Thomas is with New Jersey Medical School, Rutgers University. Michael Glick is with the School of Dental Medicine, University at Buffalo. Mel L. Kantor was with the College of Dentistry and College of Public Health, University of Kentucky when the study was done and is currently with the Institute of Health Sciences, University of Wisconsin-Eau Claire. Received: 7/28/2014; accepted: 1/26/2015. doi: 10.1111/jphd.12093
Abstract Objectives: We assessed primary care physicians’ attitudes toward medical screening in a dental setting. Methods: A 5-point Likert scale (1 = very important/willing, 5 = very unimportant/ unwilling) survey was mailed to a nationwide sample of primary care physicians in the United States. Descriptive statistics were used for all questions, and the Friedman nonparametric analysis of variance was used for multipart questions. Results: Of 1,508 respondents, the majority felt it was valuable for dentists to conduct screening for cardiovascular disease (61 percent), hypertension (77 percent), diabetes mellitus (71 percent), and HIV infection (64 percent). Respondents were willing to discuss results with the dentist (76 percent), accept patient referrals (89 percent), and felt it was unimportant that the medical referral came from a dentist rather than a physician (52 percent). The most important consideration was patient willingness (mean rank 2.55), and the least important was duplication of roles (mean rank 3.52). Level of dentist’s training was significantly (P < 0.05) more important than duplication of roles and reimbursement (mean ranks 2.84, 3.52, and 3.14, respectively), and significantly less important than patient willingness (mean rank 2.55). Conclusions: Primary care physicians considered chairside medical screening in a dental setting to be valuable and worthwhile.
Journal of Public Health Dentistry 75 (2015) 225–233
Introduction The 2010 Patient Protection and Affordable Care Act aims to slow increasing health care costs while improving the health care delivery system with a strong emphasis on prevention and primary care (1). Hence, identifying patients at risk of developing disease or with early stage disease will become increasingly important. There is also a growing appreciation for the value of an integrated approach to health care (the health home concept) and the integration across health care disciplines (2-5). There has long been a push to involve physicians in basic oral health prevention activities for children, such as placing sealants and fluoride varnishes. At the other © 2015 American Association of Public Health Dentistry
end of the spectrum is the integration of dentists into strategies to prevent and control medical conditions of public health significance. This is consistent with an expanded role for dentists that was suggested more than a decade ago and is increasingly gaining traction (2,3). The integration of dentists and physicians in patient-centered health care services also fits in with the health home concept (4,5). There is a growing body of research literature demonstrating the efficacy and potential yield of medical screening in a dental setting to identify at-risk asymptomatic patients who are unaware of their risk of developing coronary heart disease and diabetes, and who could benefit from strategies to prevent disease onset or control disease severity (6-10). 225
Physicians’ attitudes toward medical screening in a dental setting
Dentists and patients alike have a favorable attitude toward point-of-care medical screening in the dental office; dentists are willing to incorporate it into their practice and patients are willing to participate in chairside medical screening (1113). A 2008 national survey of practicing general dentists found that the majority thought it was important for dentists to screen for HIV, hepatitis, diabetes, cardiovascular disease, and hypertension, and were willing to do so and refer patients for medical follow-up (11). Another US study conducted since obtained similar results and also documented a willingness to address medical conditions such as obesity (12). A survey among adult dental patients found that the majority felt it was important for dentists to conduct medical screening that yields immediate results, would be willing to have a dentist discuss results during the visit, and receive a referral to a physician (13). Confidentiality was their most important concern, and the fact that the screening was being done by a dentist and not by a physician was the least important. The full benefit of chairside medical screening in a dental setting will necessitate collaboration with primary care physicians. A small local survey of primary care physicians conducted in the late 1980s revealed that physicians of that era had mixed reactions to dentists engaging in cardiovascular risk factor screening (14). To date, no national study has evaluated the attitudes of primary care physicians on this expanded role for dentists. The current study was designed to address this issue.
Methods The study took place February to April 2012. A selfadministered questionnaire was mailed to a stratified random sample of 6,000 allopathic (Doctor of Medicine, MD) and 1,500 osteopathic (Doctor of Osteopathic Medicine, DO) primary care physicians (family medicine and internal medicine) practicing in the United States. The sampling frame was the American Medical Association Physician List, a master file of all US-based practicing physicians (Medical Marketing Service Inc.). The questionnaire included eight ordinal scale questions, each consisting of a series of related items that addressed physicians’ attitudes regarding screening for medical conditions in a dental setting. The response scale ranged from 1 being the most positive (very effective, very important, or very willing) to 5 being the most negative (not at all effective, very unimportant, or very unwilling). Demographic information included gender, age group, race/ethnicity, medical degree type, medical specialty, years in practice, and practice location. The survey was piloted among a convenience sample of physicians for readability, clarity, and consistency. The survey was conducted following Dillman’s “Tailored Design Method” and included up to four contacts with each 226
B.L. Greenberg et al.
respondent (15). All cover letters and thank you postcards were personalized with the respondent’s name and address. To assure the respondent’s anonymity, there were no identification numbers on the questionnaires. To track nonresponders and allow for targeted, cost-effective, follow-up mailing, a separate numbered postcard was provided and returned separately. The initial mailing included the personalized cover letter explaining the study, the questionnaire, a stamped addressed return envelope, and the numbered response postcard. A thank you/reminder postcard was sent about a week later. The second mailing took place 3 weeks later, and the third and final mailing took place 4 weeks later. Returned response postcards were used to remove first-round and second-round respondents from subsequent mailings.
Statistical analysis The distribution of responses and mean ranks were calculated. The Friedman two-way nonparametric analysis of variance (ANOVA) was conducted to test if the distribution of ranks for each of the related items in a given question was different than would be expected by chance. When appropriate, post-hoc pairwise comparisons were conducted to identify the differences using the method of Siegel and Castellan which controls for the overall type I error rate for multiple comparisons (16). Logistic regression analyses were conducted to assess whether demographic factors (gender, race/ ethnicity, age group, degree, specialty, and years in practice) were independently associated with a favorable response. A favorable response was defined as a score of 1 or 2 for questions 5, 6, and 8 and a score of 4 or 5 for question 7. Analyses were conducted using sas 9.0 (SAS Software, Cary, NC, USA); pairwise comparisons were calculated manually. P < 0.05 was considered significant. The study was approved by the University of Medicine and Dentistry of New Jersey Institutional Review Board; completion and return of the questionnaire was indicative of consent to participate.
Results Of the 7,500 potential respondents, 646 were returned due to incorrect address or no longer practicing; 1,508 respondents returned the completed forms for an effective response rate of 22 percent and a margin of error of ± 2.5 percent. Table 1 shows the demographics of the respondents: 63 percent were male, 80 percent were 31-60 years of age, 75 percent were white, 79 percent were allopathic physicians, 57 percent were family medicine practitioners, and 82 percent were practicing >10 years. The sampling pool consisted of 80 percent MDs (n = 6,000) and 20 percent DOs (n = 1,500); the final study sample consisted of 79 percent MDs and 21 percent DOs. Respondents were representative of the © 2015 American Association of Public Health Dentistry
B.L. Greenberg et al.
Physicians’ attitudes toward medical screening in a dental setting
Table 1 Demographic Characteristics of Respondents (n = 1,508) and Study Sample (n = 7,500) Demographic characteristic
Respondent no. (%)
Gender (n = 1,494)* Male 942 (63.1)† Female 552 (36.9) Age (n = 1,491) 20-30 13 (0.9) 31-40 263 (17.6) 41-50 428 (28.7) 51-60 495 (33.2) 61-70 251 (16.8) ≥71 41 (2.8) Race (n = 1,464) American Indian or Alaskan 6 (0.4) Native Native Hawaiian or Pacific 7 (0.5) Islander Asian 220 (15.0) Black or African American 68 (4.6) White 1,093 (74.7) Other 70 (4.8) Ethnicity (n = 1,315) Hispanic or Latino 76 (5.8) Not Hispanic or Latino 1,239 (94.2) Year obtain medical degree (n = 1,457) 1930-1939‡ 1 (0.1) 1940-1949 0 (0.0) 1950-1959 0 (0.0) 1960-1969 50 (3.4) 1970-1979 268 (18.4) 1980-1989 443 (30.4) 1990-1999 452 (31.0) 2000-2009 242 (16.6) 2010-2012‡ 1 (0.07) Specify degree (n = 1,489) MD 1,181 (79.3) DO 308 (20.7) Specify specialty (n = 1,457) Family medicine 835 (57.3) Internal medicine 573 (39.3) Other‡ 49 (3.4) Locale (n = 1,481) Urban 473 (31.9) Suburban 679 (45.9) Rural 329 (22.2) Years of practicing (n = 1,464) ≤10 years 270 (18.4) >10 years 1,194 (81.6)
Study sample no. (%)
Chi-squared (P)
4,814 (64.2) 2,686 (35.8)
2 (0.16)
20 (0.3) 1,459 (19.5) 2,291 (30.5) 2,354 (31.4) 1,159 (15.5) 214 (2.9)
24 (0.24)
NA NA NA NA NA NA NA NA 0 (0.0) 1 (0.01) 3 (0.04) 265 (3.5) 1,152 (15.4) 2,118 (28.2) 2,495 (33.3) 1,466 (19.5) 0 (0.0)
35 (0.24)
6,000 (80.0) 1,500 (20.0) 3,966 (52.9) 3,534 (47.1) NA
6 (0.20)
NA NA NA NA NA
* Number of respondents answering question. † Percentages are based on nonmissing frequencies. Some percentages do not add to 100% due to rounding. ‡ Response values not in the original sampling frame may represent a mismark by the respondent or an error or dated information in the American Medical Association physician list. The number of these mismatches was very small and did not influence the results. NA, not available; MD, Doctor of Medicine; DO, Doctor of Osteopathic Medicine.
© 2015 American Association of Public Health Dentistry
227
Physicians’ attitudes toward medical screening in a dental setting
population of primary care providers with regard to gender, age, year of graduation, type of medical degree, and type of specialty. Table 2 shows the distribution of responses, mean ranks, Friedman nonparametric anova, and pairwise comparisons for multipart questions 1-4. Highlights of the response distribution for each question are presented below. The percentages for the two positive categories were pooled and are presented (e.g.,“effective” corresponds to “very effective” plus “somewhat effective”).
Question 1 Ninety-two percent (91.6 percent) felt it is effective to conduct point-of-care screening for hypertension, 86.2 percent for diabetes, 76.3 percent for hyperlipidemia, and 66.7 percent for HIV infection. Hypertension was ranked most positively (2.13) followed by diabetes mellitus (2.33), and testing for HIV was ranked least positively (2.84). Pairwise comparisons showed that respondents felt that point-of-care testing for hypertension was significantly more effective than testing for all other conditions listed, and pointof-care testing for HIV infection was significantly less effective than for all other conditions.
Question 2 Ninety-five percent (95.3 percent) felt that screening tests to identify patients with or at risk of disease were worthwhile for blood pressure; 88.0 percent for height and weight; 82.3 percent for hemoglobin A1c; 80.3 percent for total cholesterol and high density lipoprotein cholesterol; and 58.0 percent for salivary HIV testing. Blood pressure was ranked most positively (2.31) followed by height and weight (2.88). Pairwise comparisons showed that respondents felt that blood pressure screening was significantly more worthwhile than all other screening tests specified. Respondents also felt that measuring height and weight was significantly more worthwhile than cholesterol and HIV testing. Respondents felt that salivary HIV screening was significantly less worthwhile compared with all other screening tests noted.
Question 3 Seventy-seven percent (77.1 percent) felt it was valuable for dentists to use screening tests with follow-up medical referral for hypertension; 71.4 percent for diabetes mellitus; 64.2 percent for HIV infection; and 60.8 percent for cardiovascular disease. Screening for hypertension was ranked most valuable (2.26) followed by screening for diabetes (2.44). Screening for cardiovascular disease was ranked least worthwhile (2.77). Pairwise comparisons showed that respondents felt that screening for hypertension was significantly more 228
B.L. Greenberg et al.
worthwhile than screening for all other conditions noted and that screening for cardiovascular disease was significantly less worthwhile than screening for all other conditions noted.
Question 4 Eighty-three percent (83.4 percent) felt patient willingness was important; 77.0 percent felt both the level of training of the dentist and their own capacity to accept referrals were important; 66.4 percent felt duplication of provider reimbursement was important; and 58.0 percent felt duplication of roles was important. Patient willingness was ranked the most important (2.55) followed by training level of the dentist (2.84) and capacity to accept referrals (2.95). Pairwise comparisons showed that patient willingness was significantly more important than all other issues noted and duplication of roles significantly less important than all other issues noted.
Questions 5-8 The response distributions for questions 5-8 are shown in Table 3. Seventy-six percent (75.7 percent) of the respondents were willing to discuss the results with the dentist (Q5). Eighty-nine percent (89.2 percent) would be willing to accept a medical referral from the dentist (Q6). Only 29.8 percent felt it was important that the referral did not come from a physician, whereas the majority of respondents (51.6 percent) felt it was not important that the referral came from a dentist rather than a physician (Q7). Forty-seven percent (46.9 percent) of the respondents said their opinion of the dentist would improve, and 44.4 percent said there would be no change (Q8). Logistic regression analyses showed that no particular demographic factor was a consistent independent predictor (adjusting for all other demographic factors) for a positive response to questions 5, 6, 7, and 8 (Table 4).
Discussion This is the first nationwide study to evaluate physicians’ attitudes toward dentists incorporating chairside medical screening in their practice. The majority of the respondents felt it was valuable for a dentist to conduct screenings with referral for medical follow-up as appropriate. Respondents felt that screening for hypertension was significantly more worthwhile than all other screening tests specified. The majority was willing to discuss the results with the dentists and accept patient referrals from a dentist. The fact that the referral did not come from a physician was unimportant, whereas the most important consideration was patient willingness followed by level of training of the dentist. © 2015 American Association of Public Health Dentistry
B.L. Greenberg et al.
Physicians’ attitudes toward medical screening in a dental setting
Table 2 Distribution of Responses, Mean Rank, and Pairwise Comparisons for Questions 1-4 1. How effective do you think it is to conduct point-of-care screening for each of the following conditions?
Very effective (1) n (%)†
Somewhat effective (2) n (%)
Not sure (3) n (%)
Somewhat ineffective (4) n (%)
Very ineffective (5) n (%)
Total
Mean rank
(a) (b) (c) (d)
1,024 (68.8) 800 (59.4) 710 (47.8) 620 (42.0)
340 (22.8) 397 (26.8) 424 (28.5) 365 (24.7)
57 (3.8) 92 (6.2) 152 (10.2) 315 (21.4)
46 (3.1) 83 (5.6) 140 (9.4) 129 (8.7)
22 (1.5) 29 (2.0) 60 (4.0) 46 (3.1)
1,489 1,481 1,486 1,475
2.13 2.33 2.69 2.84
Hypertension Diabetes mellitus Hyperlipidemia HIV infection
Pairwise comparisons‡ a a b c d
b
c
d
*
* *
* * *
n = 1,463; Friedman nonparametric analysis of variance = 658; degrees of freedom = 3; P < 0.001. * Significantly different at P < 0.05. ‡ Critical value = 0.1259. 2. Consider the following point-of-care screening tests that yield immediate results. How worthwhile is each one for identifying patients with disease or at risk of developing disease?
Very worthwhile (1) n (%)†
Somewhat worthwhile (2) n (%)
Not sure (3) n (%)
Not very worthwhile (4) n (%)
Not at all worthwhile (5) n (%)
Total
Mean rank
(a) (b) (c) (d) (e)
1,191 (79.7) 867 (58.2) 809 (54.5) 744 (50.2) 513 (34.7)
233 (15.6) 444 (29.8) 412 (27.8) 447 (30.1) 347 (23.3)
26 (1.7) 76 (5.1) 112 (7.5) 113 (8.0) 304 (20.4)
27 (1.8) 72 (4.8) 106 (7.1) 121 (8.2) 135 (9.1)
18 (1.2) 30 (2.0) 45 (3.0) 52 (3.7) 94 (6.3)
1,495 1,489 1,483 1,483 1,489
2.31 2.88 2.99 3.14 3.68
Blood pressure Height and weight Hemoglobin A1c Total cholesterol and HDL Salivary HIV testing
Pairwise comparisons‡ a a b c d e
b
c
d
e
*
*
* *
* * * *
n = 1,458; Friedman nonparametric analysis of variance = 1,026; degrees of freedom = 4; P < 0.001. * Significantly different at P < 0.05. ‡ Critical value = 0.1644. 3. Consider a dentist uses the above noted point-of-care tests in the dental setting. How valuable do you think it is for a dentist to screen for each of the following conditions with referral for medical follow-up as needed?
Very valuable (1) n (%)†
Somewhat valuable (2) n (%)
Not sure (3) n (%)
Not very valuable (4) n (%)
Not at all valuable (5) n (%)
Total
Mean rank
(a) (b) (c) (d)
699 (46.7) 597 (40.0) 609 (40.9) 456 (30.6)
455 (30.4) 468 (31.4) 347 (23.3) 451 (30.2)
148 (9.9) 189 (12.7) 304 (20.4) 257 (17.2)
116 (7.7) 142 (9.5) 136 (9.1) 215 (14.4)
80 (5.3) 96 (6.4) 94 (6.3) 112 (7.5)
1,498 1,492 1,489 1,491
2.26 2.44 2.52 2.77
Hypertension Diabetes mellitus HIV infection Cardiovascular diseases
Pairwise comparisons‡ a a b c d
b
c
d
*
*
* * *
n = 1,481; Friedman nonparametric analysis of variance = 323; degrees of freedom = 3; P < 0.001. * Significantly different at P < 0.05. ‡ Critical value = 0.1252. 4. If a proposed strategy to control disease morbidity/mortality included point-of-care medical screening by dentists with referral to a physician for medical follow-up, how important would each of the following be to you?
Very important (1) n (%)†
Somewhat important (2) n (%)
Not sure (3) n (%)
Not very worthwhile (4) n (%)
(a) (b) (c) (d)
807 (54.3) 690 (46.5) 594 (40.1) 573 (38.7)
432 (29.1) 452 (30.5) 546 (36.9) 411 (27.7)
181 (12.2) 141 (9.5) 152 (10.3) 279 (18.8)
425 (28.7)
433 (29.3)
279 (18.9)
Patient willingness Level of training of the dentist Capacity to accept referrals Duplication of provider reimbursement for confirmatory testing by physician (e) Duplication of roles by dentist and physician
Not at all worthwhile (5) n (%)
Total
Mean rank
41 (2.8) 122 (8.2) 121 (8.2) 138 (9.3)
26 (1.7) 79 (5.3) 68 (4.6) 81 (5.5)
1,487 1,484 1,481 1,487
2.55 2.84 2.95 3.14
a b c d
224 (15.1)
118 (8.0)
1,479
3.52
e
Pairwise comparisons‡ a
b
c
d
e
*
*
* * *
* * * *
n = 1,467; Friedman nonparametric analysis of variance = 476; degrees of freedom = 4; P < 0.001. * Significantly different at P < 0.05. ‡ Critical value = 0.1639. † Percentages are based on nonmissing frequencies. ‡ Critical value are based on the number of comparisons and the sample size for the specific question. HDL, high density lipoprotein.
© 2015 American Association of Public Health Dentistry
229
Physicians’ attitudes toward medical screening in a dental setting
B.L. Greenberg et al.
Table 3 Distribution of Responses for Questions 5-8 5. How willing would you be to discuss with the dentist the results of a point-of-care screening for medical conditions conducted in a dental setting?
6. If a dentist conducts point-of-care screening for medical conditions that yields immediate results, how willing would you be to accept a patient medical referral from a dentist?
7. How important would it be that the medical referral came from a dentist and not a physician?
8. If a dentist conducts point-of-care medical screening in the dental office, how would your professional opinion about the dentist change?
Very willing (1) n (%)*
Somewhat willing (2) n (%)
Not sure (3) n (%)
Somewhat unwilling (4) n (%)
Very unwilling (5) n (%)
Total
763 (51.0)
369 (24.7)
164 (11.0)
121 (8.1)
78 (5.2)
1,496
Very willing (1) n (%)*
Somewhat willing (2) n (%)
Not sure (3) n (%)
Somewhat unwilling (4) n (%)
Very unwilling (5) n (%)
Total
1,056 (70.7)
276 (18.5)
85 (5.7)
37 (2.5)
39 (2.6)
1,493
Very important (1) n (%)*
Somewhat important (2) n (%)
Not sure (3) n (%)
Somewhat unimportant (4) n (%)
Very unimportant (5) n (%)
Total
212 (14.2)
233 (15.6)
277 (18.5)
347 (23.2)
425 (28.4)
1,484
Improve a lot (1) n (%)*
Improve somewhat (2) n (%)
No change (3) n (%)
Decline somewhat (4) n (%)
Decline a lot (5) n (%)
Total
282 (19.0)
414 (27.9)
659 (44.4)
78 (5.3)
51 (3.4)
1,484
* Percentages are based on nonmissing frequencies.
Less experienced practitioners (practicing 10 years (ref)
OR
Q5: Willing to discuss results with dentist
Table 4 Summary of Logistic Regression Analyses for Positive Responses to Questions 5-8
1.18
0.84 1.03
1.26
1.55 1.90 0.81
0.54 0.66
0.93
OR
0.79-1.77
0.67-1.07 0.55-1.90
0.96-1.67
0.91-2.63 1.38-2.63 0.49-1.32
0.33-0.88 0.49-0.88
0.74-1.18
95% CI
0.412
0.153 0.932
0.101
0.104
Physicians’ attitudes toward medical screening in a dental setting Barbara L. Greenberg, MSc, PhD1; Pauline A. Thomas, MD2; Michael Glick, DMD3; Mel L. Kantor, DDS, MPH, PhD4 1 2 3 4
School of Health Sciences and Practice New York Medical College, Valhalla, NY, USA New Jersey Medical School, Rutgers University, Newark, NJ, USA School of Dental Medicine, University at Buffalo, Buffalo, NY, USA College of Dentistry and College of Public Health, University of Kentucky, Lexington, KY, USA
Keywords attitudes; screening; diabetes; heart disease. Correspondence Dr. Barbara L. Greenberg, School of Health Sciences and Practice New York Medical College, 40 Sunshine Cottage Road, Valhalla, NY 10595. Tel.: 914-594-3635; Fax: 914-594-4853; e-mail: barbara_greenberg @nymc.edu. Barbara L. Greenberg is with the School of Health Sciences and Practice New York Medical College. Pauline A. Thomas is with New Jersey Medical School, Rutgers University. Michael Glick is with the School of Dental Medicine, University at Buffalo. Mel L. Kantor was with the College of Dentistry and College of Public Health, University of Kentucky when the study was done and is currently with the Institute of Health Sciences, University of Wisconsin-Eau Claire. Received: 7/28/2014; accepted: 1/26/2015. doi: 10.1111/jphd.12093
Abstract Objectives: We assessed primary care physicians’ attitudes toward medical screening in a dental setting. Methods: A 5-point Likert scale (1 = very important/willing, 5 = very unimportant/ unwilling) survey was mailed to a nationwide sample of primary care physicians in the United States. Descriptive statistics were used for all questions, and the Friedman nonparametric analysis of variance was used for multipart questions. Results: Of 1,508 respondents, the majority felt it was valuable for dentists to conduct screening for cardiovascular disease (61 percent), hypertension (77 percent), diabetes mellitus (71 percent), and HIV infection (64 percent). Respondents were willing to discuss results with the dentist (76 percent), accept patient referrals (89 percent), and felt it was unimportant that the medical referral came from a dentist rather than a physician (52 percent). The most important consideration was patient willingness (mean rank 2.55), and the least important was duplication of roles (mean rank 3.52). Level of dentist’s training was significantly (P < 0.05) more important than duplication of roles and reimbursement (mean ranks 2.84, 3.52, and 3.14, respectively), and significantly less important than patient willingness (mean rank 2.55). Conclusions: Primary care physicians considered chairside medical screening in a dental setting to be valuable and worthwhile.
Journal of Public Health Dentistry 75 (2015) 225–233
Introduction The 2010 Patient Protection and Affordable Care Act aims to slow increasing health care costs while improving the health care delivery system with a strong emphasis on prevention and primary care (1). Hence, identifying patients at risk of developing disease or with early stage disease will become increasingly important. There is also a growing appreciation for the value of an integrated approach to health care (the health home concept) and the integration across health care disciplines (2-5). There has long been a push to involve physicians in basic oral health prevention activities for children, such as placing sealants and fluoride varnishes. At the other © 2015 American Association of Public Health Dentistry
end of the spectrum is the integration of dentists into strategies to prevent and control medical conditions of public health significance. This is consistent with an expanded role for dentists that was suggested more than a decade ago and is increasingly gaining traction (2,3). The integration of dentists and physicians in patient-centered health care services also fits in with the health home concept (4,5). There is a growing body of research literature demonstrating the efficacy and potential yield of medical screening in a dental setting to identify at-risk asymptomatic patients who are unaware of their risk of developing coronary heart disease and diabetes, and who could benefit from strategies to prevent disease onset or control disease severity (6-10). 225
Physicians’ attitudes toward medical screening in a dental setting
Dentists and patients alike have a favorable attitude toward point-of-care medical screening in the dental office; dentists are willing to incorporate it into their practice and patients are willing to participate in chairside medical screening (1113). A 2008 national survey of practicing general dentists found that the majority thought it was important for dentists to screen for HIV, hepatitis, diabetes, cardiovascular disease, and hypertension, and were willing to do so and refer patients for medical follow-up (11). Another US study conducted since obtained similar results and also documented a willingness to address medical conditions such as obesity (12). A survey among adult dental patients found that the majority felt it was important for dentists to conduct medical screening that yields immediate results, would be willing to have a dentist discuss results during the visit, and receive a referral to a physician (13). Confidentiality was their most important concern, and the fact that the screening was being done by a dentist and not by a physician was the least important. The full benefit of chairside medical screening in a dental setting will necessitate collaboration with primary care physicians. A small local survey of primary care physicians conducted in the late 1980s revealed that physicians of that era had mixed reactions to dentists engaging in cardiovascular risk factor screening (14). To date, no national study has evaluated the attitudes of primary care physicians on this expanded role for dentists. The current study was designed to address this issue.
Methods The study took place February to April 2012. A selfadministered questionnaire was mailed to a stratified random sample of 6,000 allopathic (Doctor of Medicine, MD) and 1,500 osteopathic (Doctor of Osteopathic Medicine, DO) primary care physicians (family medicine and internal medicine) practicing in the United States. The sampling frame was the American Medical Association Physician List, a master file of all US-based practicing physicians (Medical Marketing Service Inc.). The questionnaire included eight ordinal scale questions, each consisting of a series of related items that addressed physicians’ attitudes regarding screening for medical conditions in a dental setting. The response scale ranged from 1 being the most positive (very effective, very important, or very willing) to 5 being the most negative (not at all effective, very unimportant, or very unwilling). Demographic information included gender, age group, race/ethnicity, medical degree type, medical specialty, years in practice, and practice location. The survey was piloted among a convenience sample of physicians for readability, clarity, and consistency. The survey was conducted following Dillman’s “Tailored Design Method” and included up to four contacts with each 226
B.L. Greenberg et al.
respondent (15). All cover letters and thank you postcards were personalized with the respondent’s name and address. To assure the respondent’s anonymity, there were no identification numbers on the questionnaires. To track nonresponders and allow for targeted, cost-effective, follow-up mailing, a separate numbered postcard was provided and returned separately. The initial mailing included the personalized cover letter explaining the study, the questionnaire, a stamped addressed return envelope, and the numbered response postcard. A thank you/reminder postcard was sent about a week later. The second mailing took place 3 weeks later, and the third and final mailing took place 4 weeks later. Returned response postcards were used to remove first-round and second-round respondents from subsequent mailings.
Statistical analysis The distribution of responses and mean ranks were calculated. The Friedman two-way nonparametric analysis of variance (ANOVA) was conducted to test if the distribution of ranks for each of the related items in a given question was different than would be expected by chance. When appropriate, post-hoc pairwise comparisons were conducted to identify the differences using the method of Siegel and Castellan which controls for the overall type I error rate for multiple comparisons (16). Logistic regression analyses were conducted to assess whether demographic factors (gender, race/ ethnicity, age group, degree, specialty, and years in practice) were independently associated with a favorable response. A favorable response was defined as a score of 1 or 2 for questions 5, 6, and 8 and a score of 4 or 5 for question 7. Analyses were conducted using sas 9.0 (SAS Software, Cary, NC, USA); pairwise comparisons were calculated manually. P < 0.05 was considered significant. The study was approved by the University of Medicine and Dentistry of New Jersey Institutional Review Board; completion and return of the questionnaire was indicative of consent to participate.
Results Of the 7,500 potential respondents, 646 were returned due to incorrect address or no longer practicing; 1,508 respondents returned the completed forms for an effective response rate of 22 percent and a margin of error of ± 2.5 percent. Table 1 shows the demographics of the respondents: 63 percent were male, 80 percent were 31-60 years of age, 75 percent were white, 79 percent were allopathic physicians, 57 percent were family medicine practitioners, and 82 percent were practicing >10 years. The sampling pool consisted of 80 percent MDs (n = 6,000) and 20 percent DOs (n = 1,500); the final study sample consisted of 79 percent MDs and 21 percent DOs. Respondents were representative of the © 2015 American Association of Public Health Dentistry
B.L. Greenberg et al.
Physicians’ attitudes toward medical screening in a dental setting
Table 1 Demographic Characteristics of Respondents (n = 1,508) and Study Sample (n = 7,500) Demographic characteristic
Respondent no. (%)
Gender (n = 1,494)* Male 942 (63.1)† Female 552 (36.9) Age (n = 1,491) 20-30 13 (0.9) 31-40 263 (17.6) 41-50 428 (28.7) 51-60 495 (33.2) 61-70 251 (16.8) ≥71 41 (2.8) Race (n = 1,464) American Indian or Alaskan 6 (0.4) Native Native Hawaiian or Pacific 7 (0.5) Islander Asian 220 (15.0) Black or African American 68 (4.6) White 1,093 (74.7) Other 70 (4.8) Ethnicity (n = 1,315) Hispanic or Latino 76 (5.8) Not Hispanic or Latino 1,239 (94.2) Year obtain medical degree (n = 1,457) 1930-1939‡ 1 (0.1) 1940-1949 0 (0.0) 1950-1959 0 (0.0) 1960-1969 50 (3.4) 1970-1979 268 (18.4) 1980-1989 443 (30.4) 1990-1999 452 (31.0) 2000-2009 242 (16.6) 2010-2012‡ 1 (0.07) Specify degree (n = 1,489) MD 1,181 (79.3) DO 308 (20.7) Specify specialty (n = 1,457) Family medicine 835 (57.3) Internal medicine 573 (39.3) Other‡ 49 (3.4) Locale (n = 1,481) Urban 473 (31.9) Suburban 679 (45.9) Rural 329 (22.2) Years of practicing (n = 1,464) ≤10 years 270 (18.4) >10 years 1,194 (81.6)
Study sample no. (%)
Chi-squared (P)
4,814 (64.2) 2,686 (35.8)
2 (0.16)
20 (0.3) 1,459 (19.5) 2,291 (30.5) 2,354 (31.4) 1,159 (15.5) 214 (2.9)
24 (0.24)
NA NA NA NA NA NA NA NA 0 (0.0) 1 (0.01) 3 (0.04) 265 (3.5) 1,152 (15.4) 2,118 (28.2) 2,495 (33.3) 1,466 (19.5) 0 (0.0)
35 (0.24)
6,000 (80.0) 1,500 (20.0) 3,966 (52.9) 3,534 (47.1) NA
6 (0.20)
NA NA NA NA NA
* Number of respondents answering question. † Percentages are based on nonmissing frequencies. Some percentages do not add to 100% due to rounding. ‡ Response values not in the original sampling frame may represent a mismark by the respondent or an error or dated information in the American Medical Association physician list. The number of these mismatches was very small and did not influence the results. NA, not available; MD, Doctor of Medicine; DO, Doctor of Osteopathic Medicine.
© 2015 American Association of Public Health Dentistry
227
Physicians’ attitudes toward medical screening in a dental setting
population of primary care providers with regard to gender, age, year of graduation, type of medical degree, and type of specialty. Table 2 shows the distribution of responses, mean ranks, Friedman nonparametric anova, and pairwise comparisons for multipart questions 1-4. Highlights of the response distribution for each question are presented below. The percentages for the two positive categories were pooled and are presented (e.g.,“effective” corresponds to “very effective” plus “somewhat effective”).
Question 1 Ninety-two percent (91.6 percent) felt it is effective to conduct point-of-care screening for hypertension, 86.2 percent for diabetes, 76.3 percent for hyperlipidemia, and 66.7 percent for HIV infection. Hypertension was ranked most positively (2.13) followed by diabetes mellitus (2.33), and testing for HIV was ranked least positively (2.84). Pairwise comparisons showed that respondents felt that point-of-care testing for hypertension was significantly more effective than testing for all other conditions listed, and pointof-care testing for HIV infection was significantly less effective than for all other conditions.
Question 2 Ninety-five percent (95.3 percent) felt that screening tests to identify patients with or at risk of disease were worthwhile for blood pressure; 88.0 percent for height and weight; 82.3 percent for hemoglobin A1c; 80.3 percent for total cholesterol and high density lipoprotein cholesterol; and 58.0 percent for salivary HIV testing. Blood pressure was ranked most positively (2.31) followed by height and weight (2.88). Pairwise comparisons showed that respondents felt that blood pressure screening was significantly more worthwhile than all other screening tests specified. Respondents also felt that measuring height and weight was significantly more worthwhile than cholesterol and HIV testing. Respondents felt that salivary HIV screening was significantly less worthwhile compared with all other screening tests noted.
Question 3 Seventy-seven percent (77.1 percent) felt it was valuable for dentists to use screening tests with follow-up medical referral for hypertension; 71.4 percent for diabetes mellitus; 64.2 percent for HIV infection; and 60.8 percent for cardiovascular disease. Screening for hypertension was ranked most valuable (2.26) followed by screening for diabetes (2.44). Screening for cardiovascular disease was ranked least worthwhile (2.77). Pairwise comparisons showed that respondents felt that screening for hypertension was significantly more 228
B.L. Greenberg et al.
worthwhile than screening for all other conditions noted and that screening for cardiovascular disease was significantly less worthwhile than screening for all other conditions noted.
Question 4 Eighty-three percent (83.4 percent) felt patient willingness was important; 77.0 percent felt both the level of training of the dentist and their own capacity to accept referrals were important; 66.4 percent felt duplication of provider reimbursement was important; and 58.0 percent felt duplication of roles was important. Patient willingness was ranked the most important (2.55) followed by training level of the dentist (2.84) and capacity to accept referrals (2.95). Pairwise comparisons showed that patient willingness was significantly more important than all other issues noted and duplication of roles significantly less important than all other issues noted.
Questions 5-8 The response distributions for questions 5-8 are shown in Table 3. Seventy-six percent (75.7 percent) of the respondents were willing to discuss the results with the dentist (Q5). Eighty-nine percent (89.2 percent) would be willing to accept a medical referral from the dentist (Q6). Only 29.8 percent felt it was important that the referral did not come from a physician, whereas the majority of respondents (51.6 percent) felt it was not important that the referral came from a dentist rather than a physician (Q7). Forty-seven percent (46.9 percent) of the respondents said their opinion of the dentist would improve, and 44.4 percent said there would be no change (Q8). Logistic regression analyses showed that no particular demographic factor was a consistent independent predictor (adjusting for all other demographic factors) for a positive response to questions 5, 6, 7, and 8 (Table 4).
Discussion This is the first nationwide study to evaluate physicians’ attitudes toward dentists incorporating chairside medical screening in their practice. The majority of the respondents felt it was valuable for a dentist to conduct screenings with referral for medical follow-up as appropriate. Respondents felt that screening for hypertension was significantly more worthwhile than all other screening tests specified. The majority was willing to discuss the results with the dentists and accept patient referrals from a dentist. The fact that the referral did not come from a physician was unimportant, whereas the most important consideration was patient willingness followed by level of training of the dentist. © 2015 American Association of Public Health Dentistry
B.L. Greenberg et al.
Physicians’ attitudes toward medical screening in a dental setting
Table 2 Distribution of Responses, Mean Rank, and Pairwise Comparisons for Questions 1-4 1. How effective do you think it is to conduct point-of-care screening for each of the following conditions?
Very effective (1) n (%)†
Somewhat effective (2) n (%)
Not sure (3) n (%)
Somewhat ineffective (4) n (%)
Very ineffective (5) n (%)
Total
Mean rank
(a) (b) (c) (d)
1,024 (68.8) 800 (59.4) 710 (47.8) 620 (42.0)
340 (22.8) 397 (26.8) 424 (28.5) 365 (24.7)
57 (3.8) 92 (6.2) 152 (10.2) 315 (21.4)
46 (3.1) 83 (5.6) 140 (9.4) 129 (8.7)
22 (1.5) 29 (2.0) 60 (4.0) 46 (3.1)
1,489 1,481 1,486 1,475
2.13 2.33 2.69 2.84
Hypertension Diabetes mellitus Hyperlipidemia HIV infection
Pairwise comparisons‡ a a b c d
b
c
d
*
* *
* * *
n = 1,463; Friedman nonparametric analysis of variance = 658; degrees of freedom = 3; P < 0.001. * Significantly different at P < 0.05. ‡ Critical value = 0.1259. 2. Consider the following point-of-care screening tests that yield immediate results. How worthwhile is each one for identifying patients with disease or at risk of developing disease?
Very worthwhile (1) n (%)†
Somewhat worthwhile (2) n (%)
Not sure (3) n (%)
Not very worthwhile (4) n (%)
Not at all worthwhile (5) n (%)
Total
Mean rank
(a) (b) (c) (d) (e)
1,191 (79.7) 867 (58.2) 809 (54.5) 744 (50.2) 513 (34.7)
233 (15.6) 444 (29.8) 412 (27.8) 447 (30.1) 347 (23.3)
26 (1.7) 76 (5.1) 112 (7.5) 113 (8.0) 304 (20.4)
27 (1.8) 72 (4.8) 106 (7.1) 121 (8.2) 135 (9.1)
18 (1.2) 30 (2.0) 45 (3.0) 52 (3.7) 94 (6.3)
1,495 1,489 1,483 1,483 1,489
2.31 2.88 2.99 3.14 3.68
Blood pressure Height and weight Hemoglobin A1c Total cholesterol and HDL Salivary HIV testing
Pairwise comparisons‡ a a b c d e
b
c
d
e
*
*
* *
* * * *
n = 1,458; Friedman nonparametric analysis of variance = 1,026; degrees of freedom = 4; P < 0.001. * Significantly different at P < 0.05. ‡ Critical value = 0.1644. 3. Consider a dentist uses the above noted point-of-care tests in the dental setting. How valuable do you think it is for a dentist to screen for each of the following conditions with referral for medical follow-up as needed?
Very valuable (1) n (%)†
Somewhat valuable (2) n (%)
Not sure (3) n (%)
Not very valuable (4) n (%)
Not at all valuable (5) n (%)
Total
Mean rank
(a) (b) (c) (d)
699 (46.7) 597 (40.0) 609 (40.9) 456 (30.6)
455 (30.4) 468 (31.4) 347 (23.3) 451 (30.2)
148 (9.9) 189 (12.7) 304 (20.4) 257 (17.2)
116 (7.7) 142 (9.5) 136 (9.1) 215 (14.4)
80 (5.3) 96 (6.4) 94 (6.3) 112 (7.5)
1,498 1,492 1,489 1,491
2.26 2.44 2.52 2.77
Hypertension Diabetes mellitus HIV infection Cardiovascular diseases
Pairwise comparisons‡ a a b c d
b
c
d
*
*
* * *
n = 1,481; Friedman nonparametric analysis of variance = 323; degrees of freedom = 3; P < 0.001. * Significantly different at P < 0.05. ‡ Critical value = 0.1252. 4. If a proposed strategy to control disease morbidity/mortality included point-of-care medical screening by dentists with referral to a physician for medical follow-up, how important would each of the following be to you?
Very important (1) n (%)†
Somewhat important (2) n (%)
Not sure (3) n (%)
Not very worthwhile (4) n (%)
(a) (b) (c) (d)
807 (54.3) 690 (46.5) 594 (40.1) 573 (38.7)
432 (29.1) 452 (30.5) 546 (36.9) 411 (27.7)
181 (12.2) 141 (9.5) 152 (10.3) 279 (18.8)
425 (28.7)
433 (29.3)
279 (18.9)
Patient willingness Level of training of the dentist Capacity to accept referrals Duplication of provider reimbursement for confirmatory testing by physician (e) Duplication of roles by dentist and physician
Not at all worthwhile (5) n (%)
Total
Mean rank
41 (2.8) 122 (8.2) 121 (8.2) 138 (9.3)
26 (1.7) 79 (5.3) 68 (4.6) 81 (5.5)
1,487 1,484 1,481 1,487
2.55 2.84 2.95 3.14
a b c d
224 (15.1)
118 (8.0)
1,479
3.52
e
Pairwise comparisons‡ a
b
c
d
e
*
*
* * *
* * * *
n = 1,467; Friedman nonparametric analysis of variance = 476; degrees of freedom = 4; P < 0.001. * Significantly different at P < 0.05. ‡ Critical value = 0.1639. † Percentages are based on nonmissing frequencies. ‡ Critical value are based on the number of comparisons and the sample size for the specific question. HDL, high density lipoprotein.
© 2015 American Association of Public Health Dentistry
229
Physicians’ attitudes toward medical screening in a dental setting
B.L. Greenberg et al.
Table 3 Distribution of Responses for Questions 5-8 5. How willing would you be to discuss with the dentist the results of a point-of-care screening for medical conditions conducted in a dental setting?
6. If a dentist conducts point-of-care screening for medical conditions that yields immediate results, how willing would you be to accept a patient medical referral from a dentist?
7. How important would it be that the medical referral came from a dentist and not a physician?
8. If a dentist conducts point-of-care medical screening in the dental office, how would your professional opinion about the dentist change?
Very willing (1) n (%)*
Somewhat willing (2) n (%)
Not sure (3) n (%)
Somewhat unwilling (4) n (%)
Very unwilling (5) n (%)
Total
763 (51.0)
369 (24.7)
164 (11.0)
121 (8.1)
78 (5.2)
1,496
Very willing (1) n (%)*
Somewhat willing (2) n (%)
Not sure (3) n (%)
Somewhat unwilling (4) n (%)
Very unwilling (5) n (%)
Total
1,056 (70.7)
276 (18.5)
85 (5.7)
37 (2.5)
39 (2.6)
1,493
Very important (1) n (%)*
Somewhat important (2) n (%)
Not sure (3) n (%)
Somewhat unimportant (4) n (%)
Very unimportant (5) n (%)
Total
212 (14.2)
233 (15.6)
277 (18.5)
347 (23.2)
425 (28.4)
1,484
Improve a lot (1) n (%)*
Improve somewhat (2) n (%)
No change (3) n (%)
Decline somewhat (4) n (%)
Decline a lot (5) n (%)
Total
282 (19.0)
414 (27.9)
659 (44.4)
78 (5.3)
51 (3.4)
1,484
* Percentages are based on nonmissing frequencies.
Less experienced practitioners (practicing 10 years (ref)
OR
Q5: Willing to discuss results with dentist
Table 4 Summary of Logistic Regression Analyses for Positive Responses to Questions 5-8
1.18
0.84 1.03
1.26
1.55 1.90 0.81
0.54 0.66
0.93
OR
0.79-1.77
0.67-1.07 0.55-1.90
0.96-1.67
0.91-2.63 1.38-2.63 0.49-1.32
0.33-0.88 0.49-0.88
0.74-1.18
95% CI
0.412
0.153 0.932
0.101
0.104
