[Skip to Content]
Sign In
Individual Sign In
Create an Account
Institutional Sign In
OpenAthens Shibboleth
[Skip to Content Landing]
Figure 1.  Notification to Referring Clinicians
Notification to Referring Clinicians

The pop-up notification occurred in real time after a clinician placed a referral order for acne in patients meeting the prespecified criteria described in the methods.

Figure 2.  Flowchart of Clinician Behavior Modification After Notification
Flowchart of Clinician Behavior Modification After Notification

BPA indicates best practice alert. The flowchart illustrates the primary outcome of behavior modification (eg, placement of premade order set or custom order set) among referring clinicians after being alerted with the notification. Of the 8 patients who did not receive treatment from the referring clinician despite having their referral to a dermatologist canceled after the notification, 2 patients were recommended to contact their local clinician for acne treatment, 1 eventually saw a dermatologist despite canceled referral, and the remaining 5 were lost to follow-up.

Table 1.  Characteristics of Patients and Clinicians Associated With Referral to a Dermatologist for Acne
Characteristics of Patients and Clinicians Associated With Referral to a Dermatologist for Acne
Table 2.  Health Care Utilization Associated With Continuing Referral to a Dermatologist for Acne Only Despite Decision Support
Health Care Utilization Associated With Continuing Referral to a Dermatologist for Acne Only Despite Decision Support
Table 3.  Comparison of Acne Treatments Initiated by Dermatologists and Referring Clinicians
Comparison of Acne Treatments Initiated by Dermatologists and Referring Clinicians
1.
Moradi Tuchayi  S, Makrantonaki  E, Ganceviciene  R, Dessinioti  C, Feldman  SR, Zouboulis  CC.  Acne vulgaris.  Nat Rev Dis Primers. 2015;1:15029. doi:10.1038/nrdp.2015.29PubMedGoogle ScholarCrossref
2.
Bhate  K, Williams  HC.  Epidemiology of acne vulgaris.  Br J Dermatol. 2013;168(3):474-485. doi:10.1111/bjd.12149PubMedGoogle ScholarCrossref
3.
Tan  JKL, Bhate  K.  A global perspective on the epidemiology of acne.  Br J Dermatol. 2015;172(suppl 1):3-12. doi:10.1111/bjd.13462PubMedGoogle ScholarCrossref
4.
Lynn  DD, Umari  T, Dunnick  CA, Dellavalle  RP.  The epidemiology of acne vulgaris in late adolescence.  Adolesc Health Med Ther. 2016;7:13-25. doi:10.2147/AHMT.S55832PubMedGoogle ScholarCrossref
5.
Tran  H, Chen  K, Lim  AC, Jabbour  J, Shumack  S.  Assessing diagnostic skill in dermatology: a comparison between general practitioners and dermatologists.  Australas J Dermatol. 2005;46(4):230-234. doi:10.1111/j.1440-0960.2005.00189.xPubMedGoogle ScholarCrossref
6.
Moreno  G, Tran  H, Chia  AL, Lim  A, Shumack  S.  Prospective study to assess general practitioners’ dermatological diagnostic skills in a referral setting.  Australas J Dermatol. 2007;48(2):77-82. doi:10.1111/j.1440-0960.2007.00340.xPubMedGoogle ScholarCrossref
7.
Liu  KJ, Hartman  RI, Joyce  C, Mostaghimi  A.  Modeling the effect of shared care to optimize acne referrals from primary care clinicians to dermatologists.  JAMA Dermatol. 2016;152(6):655-660. doi:10.1001/jamadermatol.2016.0183PubMedGoogle ScholarCrossref
8.
Faulkner  A, Mills  N, Bainton  D,  et al.  A systematic review of the effect of primary care-based service innovations on quality and patterns of referral to specialist secondary care.  Br J Gen Pract. 2003;53(496):878-884. https://www.ncbi.nlm.nih.gov/pubmed/14702909. Accessed June 3, 2018.PubMedGoogle Scholar
9.
Mallon  D, Vernacchio  L, Trudell  E,  et al.  Shared care: a quality improvement initiative to optimize primary care management of constipation.  Pediatrics. 2015;135(5):e1300-e1307. doi:10.1542/peds.2014-1962PubMedGoogle ScholarCrossref
10.
Vernacchio  L, Trudell  EK, Hresko  MT, Karlin  LI, Risko  W.  A quality improvement program to reduce unnecessary referrals for adolescent scoliosis.  Pediatrics. 2013;131(3):e912-e920. doi:10.1542/peds.2012-2359PubMedGoogle ScholarCrossref
11.
Burdick  TE, Kessler  RS.  Development and use of a clinical decision support tool for behavioral health screening in primary care clinics.  Appl Clin Inform. 2017;8(2):412-429. doi:10.4338/ACI-2016-04-RA-0068PubMedGoogle ScholarCrossref
12.
Steele  AW, Eisert  S, Witter  J,  et al.  The effect of automated alerts on provider ordering behavior in an outpatient setting.  PLoS Med. 2005;2(9):e255. doi:10.1371/journal.pmed.0020255PubMedGoogle Scholar
13.
Practitioners TRAC of G. RACGP - Acne – best practice management. https://www.racgp.org.au/afp/201009/39116. Accessed June 3, 2018.
14.
Thiboutot  D, Gollnick  H, Bettoli  V,  et al; Global Alliance to Improve Outcomes in Acne.  New insights into the management of acne: an update from the Global Alliance to Improve Outcomes in Acne group.  J Am Acad Dermatol. 2009;60(5)(suppl):S1-S50. doi:10.1016/j.jaad.2009.01.019PubMedGoogle ScholarCrossref
15.
Liao  DC.  Management of acne.  J Fam Pract. 2003;52(1):43-51. https://www.ncbi.nlm.nih.gov/pubmed/12540312. Accessed June 3, 2018.PubMedGoogle Scholar
16.
Kimball  AB, Resneck  JS  Jr.  The US dermatology workforce: a specialty remains in shortage.  J Am Acad Dermatol. 2008;59(5):741-745. doi:10.1016/j.jaad.2008.06.037PubMedGoogle ScholarCrossref
17.
Resneck  J  Jr, Kimball  AB.  The dermatology workforce shortage.  J Am Acad Dermatol. 2004;50(1):50-54. doi:10.1016/j.jaad.2003.07.001PubMedGoogle ScholarCrossref
19.
Acne vulgaris | DermNet NZ. https://www.dermnetnz.org/topics/acne-vulgaris. Accessed January 12, 2020.
20.
van der Sijs  H, Aarts  J, Vulto  A, Berg  M.  Overriding of drug safety alerts in computerized physician order entry.  J Am Med Inform Assoc. 2006;13(2):138-147. doi:10.1197/jamia.M1809PubMedGoogle ScholarCrossref
21.
Ash  JS, Berg  M, Coiera  E.  Some unintended consequences of information technology in health care: the nature of patient care information system-related errors.  J Am Med Inform Assoc. 2004;11(2):104-112. doi:10.1197/jamia.M1471PubMedGoogle ScholarCrossref
22.
Ancker  JS, Edwards  A, Nosal  S, Hauser  D, Mauer  E, Kaushal  R; with the HITEC Investigators.  Effects of workload, work complexity, and repeated alerts on alert fatigue in a clinical decision support system.  BMC Med Inform Decis Mak. 2017;17(1):36. doi:10.1186/s12911-017-0430-8PubMedGoogle ScholarCrossref
23.
Cabana  MD, Rand  CS, Powe  NR,  et al.  Why don’t physicians follow clinical practice guidelines? A framework for improvement.  JAMA. 1999;282(15):1458-1465. doi:10.1001/jama.282.15.1458PubMedGoogle ScholarCrossref
24.
Main  DS, Cohen  SJ, DiClemente  CC.  Measuring physician readiness to change cancer screening: preliminary results.  Am J Prev Med. 1995;11(1):54-58. doi:10.1016/S0749-3797(18)30501-4PubMedGoogle ScholarCrossref
25.
Grol  R.  National standard setting for quality of care in general practice: attitudes of general practitioners and response to a set of standards.  Br J Gen Pract. 1990;40(338):361-364. https://www.ncbi.nlm.nih.gov/pubmed/2265001. Accessed June 3, 2018.PubMedGoogle Scholar
26.
The Awareness-to-Adherence Model of the Steps to Clinical Gu. Medical Care. https://journals.lww.com/lww-medicalcare/Abstract/1996/09000/The_Awareness_to_Adherence_Model_of_the_Steps_to.2.aspx. Accessed June 3, 2018.
27.
Pena  S, Hill  D, Feldman  SR.  Use of topical retinoids by dermatologists and non-dermatologists in the management of acne vulgaris.  J Am Acad Dermatol. 2016;74(6):1252-1254. doi:10.1016/j.jaad.2016.01.011PubMedGoogle ScholarCrossref
28.
Mate  KS, Mehta  NP.  A two-way street: what the United States can learn from resource-limited countries to improve health care delivery and reduce costs.  Jt Comm J Qual Patient Saf. 2015;41(5):236-239. doi:10.1016/S1553-7250(15)41032-3PubMedGoogle ScholarCrossref
29.
Rush  JL, Ibrahim  J, Saul  K, Brodell  RT.  Improving Patient Safety by Combating Alert Fatigue.  J Grad Med Educ. 2016;8(4):620-621. doi:10.4300/JGME-D-16-00186.1PubMedGoogle ScholarCrossref
30.
Reducing the Safety Hazards of Monitor Alert and Alarm Fatigue | AHRQ Patient Safety Network. https://psnet.ahrq.gov/perspective/reducing-safety-hazards-monitor-alert-and-alarm-fatigue. Accessed February 1, 2020.
31.
Alarm Fatigue: A Concept Analysis | HIMSS. https://www.himss.org/alarm-fatigue-concept-analysis. Accessed February 1, 2020.
32.
Lamel  S, Chambers  CJ, Ratnarathorn  M, Armstrong  AW.  Impact of live interactive teledermatology on diagnosis, disease management, and clinical outcomes.  Arch Dermatol. 2012;148(1):61-65. doi:10.1001/archdermatol.2011.1157PubMedGoogle ScholarCrossref
33.
Lee  JJ, English  JC  III.  Teledermatology: a review and update.  Am J Clin Dermatol. 2018;19(2):253-260. doi:10.1007/s40257-017-0317-6PubMedGoogle ScholarCrossref
34.
Yentzer  BA, Hick  J, Reese  EL, Uhas  A, Feldman  SR, Balkrishnan  R.  Acne vulgaris in the United States: a descriptive epidemiology.  Cutis. 2010;86(2):94-99.PubMedGoogle Scholar
35.
Zaenglein  AL, Pathy  AL, Schlosser  BJ,  et al.  Guidelines of care for the management of acne vulgaris.  J Am Acad Dermatol. 2016;74(5):945-73.e33. doi:10.1016/j.jaad.2015.12.037PubMedGoogle ScholarCrossref
36.
Nast  A, Dréno  B, Bettoli  V,  et al.  European evidence-based (S3) guideline for the treatment of acne - update 2016 - short version.  J Eur Acad Dermatol Venereol. 2016;30(8):1261-1268. doi:10.1111/jdv.13776PubMedGoogle ScholarCrossref
Limit 200 characters
Limit 25 characters
Conflicts of Interest Disclosure

Identify all potential conflicts of interest that might be relevant to your comment.

Conflicts of interest comprise financial interests, activities, and relationships within the past 3 years including but not limited to employment, affiliation, grants or funding, consultancies, honoraria or payment, speaker's bureaus, stock ownership or options, expert testimony, royalties, donation of medical equipment, or patents planned, pending, or issued.

Err on the side of full disclosure.

If you have no conflicts of interest, check "No potential conflicts of interest" in the box below. The information will be posted with your response.

Not all submitted comments are published. Please see our commenting policy for details.

Limit 140 characters
Limit 3600 characters or approximately 600 words
    Views 640
    Citations 0
    Original Investigation
    March 4, 2020

    Evaluation of Point-of-Care Decision Support for Adult Acne Treatment by Primary Care Clinicians

    Author Affiliations
    • 1Department of Dermatology, Brigham & Women's Hospital, Harvard Medical School, Boston, Massachusetts
    • 2University of California, San Francisco School of Medicine, San Francisco
    • 3Department of Medicine, Brigham & Women’s Hospital, Harvard Medical School, Boston, Massachusetts
    • 4Brigham and Women’s Physicians Organization, Boston, Massachusetts
    • 5Loyola University, Chicago, Illinois
    JAMA Dermatol. Published online March 4, 2020. doi:10.1001/jamadermatol.2020.0135
    Key Points

    Question  What are the downstream outcomes following implementation of a real-time, electronic decision-support tool for the treatment of patients referred for acne?

    Findings  This prospective cohort study included 260 patients referred by a primary care clinician for acne. Overall, the algorithm was associated with cancellation of the initial referral in 35 of 260 (13.5%) instances and treatment initiation by the referring clinician in 51 of 260 (19.6%) instances.

    Meaning  This decision-support algorithm was associated with modestly reduced rates of acne-related referrals to dermatologists and increased likelihood of treatment initiation by the referring clinician.

    Abstract

    Importance  Acne is a common reason for referral to dermatologists from primary care clinicians. We previously modeled the impact of algorithm-based acne care in reducing dermatology referrals, missed appointments, and treatment delays.

    Objective  To prospectively evaluate the downstream outcomes following a real-time, algorithm-based electronic decision-support tool on the treatment of patients referred for acne.

    Design, Setting, and Participants  This prospective cohort study included 260 treatment-naive patients referred to a dermatologist for the chief concern of acne, as well as the referring primary care clinicians, at 33 primary care sites affiliated with Brigham and Women’s Hospital from March 2017 to March 2018.

    Interventions  We developed and implemented a decision-support tool into the electronic medical record system at an academic medical center. The algorithm identified patients referred to a dermatologist who had not previously been treated for acne and offered guideline-based recommendations for treatment via a real-time notification.

    Main Outcomes and Measures  Treatment modification by referring clinicians.

    Results  Of 260 patients referred for acne, 209 (80.4%) were women, 146 (56.1%) were non-Hispanic white, and 236 (90.8%) listed English as the preferred language. Patients had a median (quartile 1-quartile 3) age of 28.8 years (24.4-35.1 years) and 185 of 260 had private insurance (71.1%). In total, the algorithm was associated with cancellation of the initial referral in 35 of 260 (13.5%) instances and treatment initiation by the referring clinician in 51 of 260 (19.6%) instances.

    Conclusions and Relevance  This decision-support algorithm was associated with a modest reduction in rates of acne-related referrals to dermatologists, and an increased likelihood of treatment initiation by the referring clinician.

    Introduction

    Acne is a chronic inflammatory skin disease that affects more than 85% of adolescents and frequently continues to adulthood.1-4 Although there is a high diagnostic concordance for acne between dermatologists and primary care clinicians, acne continues to be a common indication for referral to a dermatologist.5,6 Although referrals to dermatologists for acne are not discouraged, there are consequences of immediate referral by primary care clinicians who do not provide treatment, including delayed treatment and loss to follow-up.7

    Previous studies have demonstrated the potential of education and treatment algorithms to reduce unnecessary referrals to specialty care in nondermatological disease areas.8-12 Importantly, these studies support the use of real-time decision-support tools to modify physician behavior and reduce unnecessary referrals, specifically for patients with pediatric scoliosis or constipation being referred to specialists by primary care clinicians.9,10 Meanwhile, decision-support tools have also been used to improve rates of health screening questions for preventative health and laboratory test orders for monitoring of adverse drug events.11,12

    In dermatology, stepwise treatment algorithms for acne have been proposed, although implementation has been a challenge.13-15 In a previous study,7 we modeled the effect of algorithm-based acne treatment by primary care clinicians, and demonstrated shared acne care had the potential to decrease the rates of unnecessary appointments, wait time for treatment, no-show rates, and downstream costs associated with acne. Despite these predictions, prospective evaluation of algorithm-based care for acne is lacking. From a societal standpoint, shared care may reduce unnecessary referrals for acne while reallocating resources to improve access to dermatologic care, which has an average wait time of 32.3 days until the first appointment.16-18

    In this study, we prospectively evaluate the association of implementation of an algorithm-based, decision-support tool with changes in referral rates and treatment of adult acne among primary care clinicians in a single health care system across 33 clinical sites.

    Methods
    Patient Population and Development of Algorithm

    We developed a point-of-care electronic health record–based algorithm designed to alert clinicians referring treatment-naive patients with acne to a dermatologist with a pop-up notification (Figure 1). To appropriately identify treatment-naive patients, the algorithm excluded patients who had a previously documented prescription acne treatment in their electronic medical record and patients who had a documented dermatology visit for acne within 1 year of the referral date.

    If a patient meeting the above criteria was referred to a dermatologist, the algorithm generated a notification alerting the clinician that the patient was being referred to a dermatology department without having trialed prescription treatments for acne. The notification provided representative photographs of acne severity (mild, moderate, severe)19 with corresponding recommendations for treatment, based on guidelines of care for the management of acne vulgaris at the time of study design in 2015.

    The clinician was then given the option of (1) keeping the referral order without initiating treatment, (2) keeping the referral order while initiating treatment (using premade order sets tailored to acne severity), (3) removing the referral order without initiating treatment, or (4) removing the referral order while initiating treatment (using the premade order set).

    Premade order sets consisted of (1) full topical regimen for patients with mild acne (tretinoin, 0.05% cream nightly and clindamycin, 1% external solution twice daily), (2) full topical plus oral antibiotic (tretinoin, 0.05% cream nightly and clindamycin, 1% external solution twice daily, and doxycycline hyclate, 100 mg twice daily) for patients with moderate acne, and (3) an accelerated 2-week referral to a dermatologist for patients with severe acne (eTable 1 in the Supplement).

    Partners Healthcare institutional review board approval was obtained for this study and written informed consent was waived because data regarding clinician behavior were collected within the electronic medical record system (EPIC, version 2015), and the remainder of clinician and patient data were examined via medical chart review as part of standard care. Data were analyzed between October and December, 2018.

    Implementation of Algorithm

    The point-of-care algorithm was implemented in EPIC for primary care sites affiliated with Brigham and Women’s Hospital (Boston, Massachusetts) and identified patient encounters associated with an ambulatory referral order to the dermatology department for acne from March 2017 through March 2018. Among patients whose clinicians received a pop-up notification, the algorithm identified and recorded the patient’s medical record number, the referring clinician’s name, and the time/date of the pop-up notification.

    Data Collection: Initial Presentation for Acne

    The medical records of all patients for whom the referring clinician received a real-time notification for acne referral were reviewed, and patient and clinician demographic information was retrieved. Each clinician’s engagement with the real-time notification (eg, ignored notification, kept referral, removed referral) was documented, and the patient’s medication list was reviewed to assess treatment rendered by the referring physician.

    Data Collection: Referral to a Dermatologist

    Among patients whose clinicians proceeded with the referral to a dermatologist for acne, patient records were reviewed to determine whether a dermatologic appointment was made after and the status of the appointment (seen, not yet seen, no-show/canceled appointment). For patients seen by a dermatologist, we documented the diagnosis made by the dermatologist and any acne medications that were used. Medical records were reviewed at 60 days after the initial referral date.

    Statistical Analysis

    Data were recorded using Research Electronic Data Capture (version 6, REDCap). All data were analyzed and reported descriptively with median (quartile 1, quartile 3) for continuous measures and counts (percentage) for categorical variables. Topical, oral, combination acne treatments, and isotretinoin as initiated by dermatologists and referring clinicians were compared using a Fisher exact test.

    Results
    Patient and Clinician Demographics

    We identified 260 unique referrals to the dermatology department for acne from across 33 different primary care sites in Massachusetts. Of these, 188 (72.3%) patients were referred for an isolated problem of acne and 72 (27.7%) were referred for acne and a concurrent dermatologic issue. Among the 260 patients, 209 (80.4%) were women, 146 (56.1%) were non-Hispanic white, and 236 (90.8%) listed English as the preferred language (Table 1). Patients had a median (Q1, Q3) age of 28.8 (24.4, 35.1) years and 185 (71.1%) had private insurance.

    At the patient encounter level, 177 (68.1%) and 34 (13.1%) patients were seen by 103 attending physicians and 30 resident physicians, respectively. The remaining 49 (18.8%) patients were seen by 28 midlevel clinicians (eg, nurse, physician assistant) and a midwife (Table 1). One hundred ninety patients (73.1%) were seen by a female clinician. These demographic findings were similar between patients referred for acne and those referred for acne plus a concurrent skin concern.

    Decision-Support Notification and Clinician Treatment of Acne

    Overall, the decision-support notification coincided with treatment initiation in 51 of 260 (19.6%) of cases. Of these, treatment was initiated using the notification’s premade order set in 16 of 51 (31.4%) instances (Figure 2).

    Among 188 of 260 patients referred for only acne, the real-time notification was associated with clinician cancellation of the initial referral to a dermatologist for acne in 34 of 188 (18.1%) cases. Of these 34 cases, the referring clinician then initiated acne treatment in 26 (76.5%) instances. Eleven of these 34 (42.3%) cases were initiated using the premade order set. In the other 154 patients whose referrals were continued despite the notification, the referring clinician then initiated acne treatment in 17 (11.0%) cases. Only 3 of these 17 (17.6%) cases were initiated via the premade order set.

    For the 72 of 260 patients referred for acne plus another skin problem, the alert was associated with clinician cancellation of the initial referral to a dermatologist for acne in 1 (1.4%) case, and treatment was provided via an order set. In the other 71 patients whose referrals were continued despite the alert, the referring clinician then initiated acne treatment in 7 (9.9%) cases, 1 (14.3%) of which was initiated on an order set.

    Health Care Utilization Among Patients With Referrals for Acne

    Of 154 patients referred for an isolated concern of acne whose referrals were continued, 121 (78.6%) had made an appointment at 60 days after the date of referral (Table 2). Of these, 20 (16.5%) were no-shows and 5 (4.1%) had not been seen at 60 days postreferral. The remaining 96 (79.3%) patients were seen by a dermatologist, of which 90 of 96 (93.5%) were diagnosed with acne (eTable 2 in the Supplement). Overall, 96 of 154 (62.3%) patients referred to a dermatologist for acne were seen at 60 days after the referral date. These findings were similar among patients who were referred to a dermatologist for acne in addition to another skin problem.

    Comparison of Treatments by Dermatologists and Referring Clinicians

    Of the 117 patients who were seen by a dermatologist and confirmed to have acne, 62 (53.0%) were prescribed at least 1 topical medication for acne (Table 3). Oral medications including oral antibiotics, spironolactone, and/or oral contraceptives were initiated in 6 (5.1%) patients, whereas combined topical and oral medications were prescribed in 35 (30.0%) patients. Isotretinoin was started in 8 (6.8%) patients. In the remaining 6 patients, 4 (3.4%) declined proposed treatment by the dermatologist and 2 (1.7%) were recommended to continue the treatment initiated by the referring clinician.

    Among 51 patients who initiated treatment recommended by the referring clinician, 33 (64.7%), 4 (7.8%), and 10 (19.6%) were prescribed topical, oral, and combined topical and oral medications. When classified by these broad categories, it appeared that treatment selections by referring clinicians did not differ significantly from those chosen by dermatologists. At 60 days after the date of referral, only 4 (7.8%) patients treated by a referring clinician received further escalation of care (eg, additional referral to a dermatologist). Of these 4 patients, 3 (75.0%) were prescribed an acne treatment on a trial basis for fewer than 2 weeks before receiving a referral to a dermatologist.

    Of 51 patients who were prescribed treatment via the referring clinician, 15 (29.4%) were eventually evaluated by a dermatologist at 60 days. Thirteen of 15 patients (86.7%) had their regimen changed by the dermatologist. The most frequent change was the addition of 1 or more topical medications such as clindamycin (6 of 13), benzoyl peroxide (4 of 13), tretinoin (4 of 13), or adapalene (2 of 13) onto the existing regimen. Medications started by the referring clinician were discontinued and replaced with alternatives in 4 of 15 patients (26.7%). In 2 of 15 cases (13.3%), the dermatologist agreed with the initial treatment regimen by the referring clinician and did not alter the regimen.

    Discussion

    In this prospective study examining the implementation of a point-of-care decision support tool for adult acne treatment, we found that real-time notification may be modestly effective in modifying clinician referral and treatment behavior. After being presented with the notification containing acne treatment guidelines, referring clinicians often chose an appropriate regimen that approximately matched treatments prescribed by dermatologists. Although these findings have demonstrated reductions in immediate referrals to dermatologists for acne, the maximal impact of the advisory tool was not obtained, as evident from high rates of referral continuation (>80%).

    The reason for this limited success is likely multifactorial. Referring clinicians may have disregarded the real-time notification because of alert fatigue, which has been shown to explain high physician override rates of decision support alerts.20-22 Furthermore, because the notification occurred after a decision was already made to refer the patient, cognitive barriers (eg, anchoring bias) may have prevented physician uptake of the notification. Likewise, some referring physicians may have placed the referral (and received a notification) after the patient visit, making them less likely to deviate from the original plan as communicated with the patient. It is also plausible that clinicians may have declined to change their existing decision-making given the inertia of conventional practice patterns, which can contribute toward nonadherence to clinical guidelines.23-26 Finally, it is possible that some patients may have declined treatment recommended by the notification and referring clinician, and that this interaction was simply not documented.

    Despite these challenges, these findings support and expand on existing data regarding the role of primary care clinicians in shared care for evaluation and treatment of acne.6,7 Importantly, we demonstrated the decision support notification to correlate with the selection of guideline-based acne treatments by referring clinicians, which were not significantly different from treatments by dermatologists (Table 3). Although 4 (7.8%) patients treated by referring clinicians received subsequent referral to a dermatologist, most of these patients were prescribed acne medications on a trial basis for fewer than 2 weeks prior to dermatology assessment.

    Early treatment initiation by referring clinicians may also address the high rates of loss to follow-up. Among patients whose referring clinicians continued with the dermatologist referral for an isolated problem of acne, only 96 of 154 (62.3%) were seen by a dermatologist at 60 days after the initial referral date. Of the 5 patients with appointments who were not seen at the time of analysis, 2 were later evaluated by a dermatologist and prescribed an oral acne treatment (doxycycline and spironolactone); the remaining 3 patients were lost to follow-up, further underscoring the potential consequences associated with immediate referral to a dermatologist for acne.7

    Although this point-of-care notification reduced rates of immediate referral to a dermatologist and improved the likelihood of treatment initiation, a comprehensive approach embedded into the training and practice of primary care clinicians may be needed to improve the efficacy of this intervention. For instance, educational initiatives for primary care clinicians to highlight the opportunities and challenges associated with shared care for relevant disease models (eg, acne) may partially overcome the cognitive overload and alert desensitization associated with electronic health record systems, while priming physicians to be more receptive to best practice notifications.27-31 Initiatives including team-based didactics and cross-training rotations may also improve clinician preparedness for initiating treatment for acne. These approaches have the added benefit of helping to maintain antibiotic stewardship among physicians who may have become reliant on the use of antibiotics to treat acne. Finally, repositioning of future decision-support notifications to appear earlier in the clinician workflow may improve patient-physician communication regarding alternatives to immediate dermatologist referral.

    Beyond this point-of-care decision support tool, other opportunities to promote shared care for the treatment of acne should be explored. Whereas our intervention focused on the use of algorithm-based notifications in the clinical workflow of primary care clinicians, another potentially viable option to optimize acne referrals and reduce time to treatment would be to use live interactive or store-and-forward teledermatology during the patient’s initial visit at the primary care office. Shared patient care among primary care clinicians and live video teledermatologists have been studied in the literature, with 1 study of 1500 live interactive teledermatology consultations for unspecified skin diseases resulting in changes to diagnosis and treatment in 69.9% and 97.7% of the time, respectively.32,33 To further reduce the rate of unnecessary referrals for acne among treatment-naive patients, it may also be worthwhile to explore the use of live interactive teledermatology to facilitate shared care for acne.

    Limitations

    These findings must be interpreted in the context of study design. Although this was a multicenter prospective study, all practice locations were in Massachusetts and thus our results may not be generalizable given regional and geographical variations in practice. Specifically, all practice locations included in the study were affiliated with a large academic medical center and results may not be generalizable to different practice settings. Similarly, generalizability may be limited largely to adult women because sex distribution among patients in our study (80.2%) is greater than the reported sex distribution among patients with acne vulgaris in the United States (65.2%).34 In addition, the median age of patients in our sample was greater than the national average for acne primarily because our outpatient dermatology clinic does not see patients younger than 16 years. Furthermore, although we could verify an association between the notification and referral cancellation, we could only establish an association between the notification and treatment initiation by referring clinicians because it is unclear whether physicians who initiated treatment would not have done so in the absence of the decision-support tool. Finally, our data do not include referrals for acne that were coded under different names (eg, facial rash).

    Regarding treatment selection, we had included topical and oral antibiotics in our premade order sets based on regional prescription preferences in efforts to have primary care clinicians adopt our algorithm; however, there exist treatments that would be more optimal and in alignment with current acne treatment guidelines.35,36 Future studies should consider a broader selection of commonly prescribed and effective treatments for acne, including benzoyl peroxide, oral contraceptives, and spironolactone while also accounting for separate pathways for male and female patients. Although these options were not included in the present study to simplify the premade order sets, they may reduce the risk of inappropriate prescribing of oral antibiotics.

    Additional efforts aimed at gathering a larger sample size would enable granular treatment comparisons between dermatologists and primary care clinicians. Increasing follow-up time to allow for certain nonhormonal treatments to take full effect before analyzing the data and implementing a patient/clinician survey to gain additional insight into the modest success of this intervention would also be of benefit.

    Conclusions

    These data suggest that this decision-support tool may be modestly effective at reducing the rates of acne-related referrals to dermatologists among treatment-naive patients, while increasing the likelihood of treatment initiation by the referring clinician. Not only were treatments by referring primary care clinicians similar to those selected by dermatologists, the rates of subsequent referral among these patients were low. Moving forward, further exploration of educational decision support in combination with more comprehensive initiatives may optimize shared care for acne.

    Back to top
    Article Information

    Corresponding Author: Arash Mostaghimi, MD, MPA, MPH, Department of Dermatology, Brigham and Women’s Hospital, 221 Longwood Ave, Boston, MA 02115 (amostaghimi@bwh.harvard.edu).

    Accepted for Publication: January 16, 2020.

    Published Online: March 4, 2020. doi:10.1001/jamadermatol.2020.0135

    Author Contributions: Drs Li and Mostaghimi had full access to all the data in the study and take responsibility for the integrity of the data and the accuracy of the data analysis.

    Study concept and design: Li, Liu, Laskowski.

    Acquisition, analysis, or interpretation of data: Li, Pournamdari, Laskowski, Joyce.

    Drafting of the manuscript: Li, Pournamdari.

    Critical revision of the manuscript for important intellectual content: Li, Liu, Laskowski, Joyce.

    Statistical analysis: Li, Joyce.

    Obtained funding: Li, Laskowski.

    Administrative, technical, or material support: Laskowski.

    Conflicts of Interest: Dr Mostaghimi is on the editorial board of JAMA Dermatology but played no role in the acceptance or publication of the article. No other conflicts are reported.

    Funding/Support: This project was supported by the National Center for Advancing Translational Sciences, National Institutes of Health, Award Number TL1TR001062 and by the Brigham and Women’s Physician Organization Brigham Care Redesign Incubator and Startup Program (BCRISP). This project was supported by the National Center for Advancing Translational Sciences, National Institutes of Health, Award Number TL1TR001062 (Dr Li) and by the Brigham and Women’s Physician Organization Brigham Care Redesign Incubator and Startup Program (BCRISP) (Dr Mostaghimi).

    Role of the Funder/Sponsor: The funding agencies had no role in the design and conduct of the study; collection, management, analysis, and interpretation of the data; preparation, review, or approval of the manuscript; and decision to submit the manuscript for publication.

    References
    1.
    Moradi Tuchayi  S, Makrantonaki  E, Ganceviciene  R, Dessinioti  C, Feldman  SR, Zouboulis  CC.  Acne vulgaris.  Nat Rev Dis Primers. 2015;1:15029. doi:10.1038/nrdp.2015.29PubMedGoogle ScholarCrossref
    2.
    Bhate  K, Williams  HC.  Epidemiology of acne vulgaris.  Br J Dermatol. 2013;168(3):474-485. doi:10.1111/bjd.12149PubMedGoogle ScholarCrossref
    3.
    Tan  JKL, Bhate  K.  A global perspective on the epidemiology of acne.  Br J Dermatol. 2015;172(suppl 1):3-12. doi:10.1111/bjd.13462PubMedGoogle ScholarCrossref
    4.
    Lynn  DD, Umari  T, Dunnick  CA, Dellavalle  RP.  The epidemiology of acne vulgaris in late adolescence.  Adolesc Health Med Ther. 2016;7:13-25. doi:10.2147/AHMT.S55832PubMedGoogle ScholarCrossref
    5.
    Tran  H, Chen  K, Lim  AC, Jabbour  J, Shumack  S.  Assessing diagnostic skill in dermatology: a comparison between general practitioners and dermatologists.  Australas J Dermatol. 2005;46(4):230-234. doi:10.1111/j.1440-0960.2005.00189.xPubMedGoogle ScholarCrossref
    6.
    Moreno  G, Tran  H, Chia  AL, Lim  A, Shumack  S.  Prospective study to assess general practitioners’ dermatological diagnostic skills in a referral setting.  Australas J Dermatol. 2007;48(2):77-82. doi:10.1111/j.1440-0960.2007.00340.xPubMedGoogle ScholarCrossref
    7.
    Liu  KJ, Hartman  RI, Joyce  C, Mostaghimi  A.  Modeling the effect of shared care to optimize acne referrals from primary care clinicians to dermatologists.  JAMA Dermatol. 2016;152(6):655-660. doi:10.1001/jamadermatol.2016.0183PubMedGoogle ScholarCrossref
    8.
    Faulkner  A, Mills  N, Bainton  D,  et al.  A systematic review of the effect of primary care-based service innovations on quality and patterns of referral to specialist secondary care.  Br J Gen Pract. 2003;53(496):878-884. https://www.ncbi.nlm.nih.gov/pubmed/14702909. Accessed June 3, 2018.PubMedGoogle Scholar
    9.
    Mallon  D, Vernacchio  L, Trudell  E,  et al.  Shared care: a quality improvement initiative to optimize primary care management of constipation.  Pediatrics. 2015;135(5):e1300-e1307. doi:10.1542/peds.2014-1962PubMedGoogle ScholarCrossref
    10.
    Vernacchio  L, Trudell  EK, Hresko  MT, Karlin  LI, Risko  W.  A quality improvement program to reduce unnecessary referrals for adolescent scoliosis.  Pediatrics. 2013;131(3):e912-e920. doi:10.1542/peds.2012-2359PubMedGoogle ScholarCrossref
    11.
    Burdick  TE, Kessler  RS.  Development and use of a clinical decision support tool for behavioral health screening in primary care clinics.  Appl Clin Inform. 2017;8(2):412-429. doi:10.4338/ACI-2016-04-RA-0068PubMedGoogle ScholarCrossref
    12.
    Steele  AW, Eisert  S, Witter  J,  et al.  The effect of automated alerts on provider ordering behavior in an outpatient setting.  PLoS Med. 2005;2(9):e255. doi:10.1371/journal.pmed.0020255PubMedGoogle Scholar
    13.
    Practitioners TRAC of G. RACGP - Acne – best practice management. https://www.racgp.org.au/afp/201009/39116. Accessed June 3, 2018.
    14.
    Thiboutot  D, Gollnick  H, Bettoli  V,  et al; Global Alliance to Improve Outcomes in Acne.  New insights into the management of acne: an update from the Global Alliance to Improve Outcomes in Acne group.  J Am Acad Dermatol. 2009;60(5)(suppl):S1-S50. doi:10.1016/j.jaad.2009.01.019PubMedGoogle ScholarCrossref
    15.
    Liao  DC.  Management of acne.  J Fam Pract. 2003;52(1):43-51. https://www.ncbi.nlm.nih.gov/pubmed/12540312. Accessed June 3, 2018.PubMedGoogle Scholar
    16.
    Kimball  AB, Resneck  JS  Jr.  The US dermatology workforce: a specialty remains in shortage.  J Am Acad Dermatol. 2008;59(5):741-745. doi:10.1016/j.jaad.2008.06.037PubMedGoogle ScholarCrossref
    17.
    Resneck  J  Jr, Kimball  AB.  The dermatology workforce shortage.  J Am Acad Dermatol. 2004;50(1):50-54. doi:10.1016/j.jaad.2003.07.001PubMedGoogle ScholarCrossref
    19.
    Acne vulgaris | DermNet NZ. https://www.dermnetnz.org/topics/acne-vulgaris. Accessed January 12, 2020.
    20.
    van der Sijs  H, Aarts  J, Vulto  A, Berg  M.  Overriding of drug safety alerts in computerized physician order entry.  J Am Med Inform Assoc. 2006;13(2):138-147. doi:10.1197/jamia.M1809PubMedGoogle ScholarCrossref
    21.
    Ash  JS, Berg  M, Coiera  E.  Some unintended consequences of information technology in health care: the nature of patient care information system-related errors.  J Am Med Inform Assoc. 2004;11(2):104-112. doi:10.1197/jamia.M1471PubMedGoogle ScholarCrossref
    22.
    Ancker  JS, Edwards  A, Nosal  S, Hauser  D, Mauer  E, Kaushal  R; with the HITEC Investigators.  Effects of workload, work complexity, and repeated alerts on alert fatigue in a clinical decision support system.  BMC Med Inform Decis Mak. 2017;17(1):36. doi:10.1186/s12911-017-0430-8PubMedGoogle ScholarCrossref
    23.
    Cabana  MD, Rand  CS, Powe  NR,  et al.  Why don’t physicians follow clinical practice guidelines? A framework for improvement.  JAMA. 1999;282(15):1458-1465. doi:10.1001/jama.282.15.1458PubMedGoogle ScholarCrossref
    24.
    Main  DS, Cohen  SJ, DiClemente  CC.  Measuring physician readiness to change cancer screening: preliminary results.  Am J Prev Med. 1995;11(1):54-58. doi:10.1016/S0749-3797(18)30501-4PubMedGoogle ScholarCrossref
    25.
    Grol  R.  National standard setting for quality of care in general practice: attitudes of general practitioners and response to a set of standards.  Br J Gen Pract. 1990;40(338):361-364. https://www.ncbi.nlm.nih.gov/pubmed/2265001. Accessed June 3, 2018.PubMedGoogle Scholar
    26.
    The Awareness-to-Adherence Model of the Steps to Clinical Gu. Medical Care. https://journals.lww.com/lww-medicalcare/Abstract/1996/09000/The_Awareness_to_Adherence_Model_of_the_Steps_to.2.aspx. Accessed June 3, 2018.
    27.
    Pena  S, Hill  D, Feldman  SR.  Use of topical retinoids by dermatologists and non-dermatologists in the management of acne vulgaris.  J Am Acad Dermatol. 2016;74(6):1252-1254. doi:10.1016/j.jaad.2016.01.011PubMedGoogle ScholarCrossref
    28.
    Mate  KS, Mehta  NP.  A two-way street: what the United States can learn from resource-limited countries to improve health care delivery and reduce costs.  Jt Comm J Qual Patient Saf. 2015;41(5):236-239. doi:10.1016/S1553-7250(15)41032-3PubMedGoogle ScholarCrossref
    29.
    Rush  JL, Ibrahim  J, Saul  K, Brodell  RT.  Improving Patient Safety by Combating Alert Fatigue.  J Grad Med Educ. 2016;8(4):620-621. doi:10.4300/JGME-D-16-00186.1PubMedGoogle ScholarCrossref
    30.
    Reducing the Safety Hazards of Monitor Alert and Alarm Fatigue | AHRQ Patient Safety Network. https://psnet.ahrq.gov/perspective/reducing-safety-hazards-monitor-alert-and-alarm-fatigue. Accessed February 1, 2020.
    31.
    Alarm Fatigue: A Concept Analysis | HIMSS. https://www.himss.org/alarm-fatigue-concept-analysis. Accessed February 1, 2020.
    32.
    Lamel  S, Chambers  CJ, Ratnarathorn  M, Armstrong  AW.  Impact of live interactive teledermatology on diagnosis, disease management, and clinical outcomes.  Arch Dermatol. 2012;148(1):61-65. doi:10.1001/archdermatol.2011.1157PubMedGoogle ScholarCrossref
    33.
    Lee  JJ, English  JC  III.  Teledermatology: a review and update.  Am J Clin Dermatol. 2018;19(2):253-260. doi:10.1007/s40257-017-0317-6PubMedGoogle ScholarCrossref
    34.
    Yentzer  BA, Hick  J, Reese  EL, Uhas  A, Feldman  SR, Balkrishnan  R.  Acne vulgaris in the United States: a descriptive epidemiology.  Cutis. 2010;86(2):94-99.PubMedGoogle Scholar
    35.
    Zaenglein  AL, Pathy  AL, Schlosser  BJ,  et al.  Guidelines of care for the management of acne vulgaris.  J Am Acad Dermatol. 2016;74(5):945-73.e33. doi:10.1016/j.jaad.2015.12.037PubMedGoogle ScholarCrossref
    36.
    Nast  A, Dréno  B, Bettoli  V,  et al.  European evidence-based (S3) guideline for the treatment of acne - update 2016 - short version.  J Eur Acad Dermatol Venereol. 2016;30(8):1261-1268. doi:10.1111/jdv.13776PubMedGoogle ScholarCrossref
    ×