To the Editor: No-shows and same-day appointment cancellations result in underutilization of clinical space and human resources, delayed medical treatment, and reduced physician productivity. Missed appointments have been correlated with patient and appointment-level factors.1 However, few studies have looked at dermatology clinics, unique in their volume of surgical, medical, and cosmetic appointments.1 Understanding missed appointments may improve health care access and maximize clinic utilization.
We conducted a case-control study of 5 adult dermatology clinics encompassing 99,476 encounters (1/1-12/31/2019): ages 15 to 104 (mean 51.4), 62% female, excluding concurrent inpatient admission. Cases were defined as no-shows or cancellations within 24 hours of appointment and controls as completed appointments. Mann-Whitney U and Pearson χ2 tests compared differences between “show” (controls) and “no-show” (cases). No-show adjusted odds ratio (AOR) was determined by multivariable, multilevel mixed effects logistic regression (Table I), accounting for variability naturally present in data (clinic location and repeated measurements of an individual). Clinic site, patient identifier, and individual encounter as the lowest level were specified with random intercepts in a nested fashion. Demographic, socioeconomic, past medical history, and administrative variables (Table II) were modeled as risk factor covariates. Variable selection used elastic net, least absolute shrinkage and selection operator, and backwards stepwise selection excluding variables P ≥ .2 and including variables P < .1 on univariate analysis.2
Table I.
Results of multivariate logistic regression (mixed effect modeling) for independent predictors of no-shows/same-day cancellations in 99,476 dermatology encounters
| Variable | AOR | 95% CI | P value |
|---|---|---|---|
| Age | 0.904 | 0.881-0.928 | <.001 |
| Male sex | 1.021 | 0.956-1.09 | .543 |
| Race (compared to White) | |||
| American Indian/Alaska Native | 1.828 | 0.918-3.642 | .086 |
| Asian or Pacific Islander | 1.032 | 0.892-1.193 | .672 |
| Black or African American | 2.099 | 1.945-2.265 | <.001 |
| Hispanic ethnicity | 1.143 | 0.973-1.342 | .104 |
| Interpreter needed | 1.136 | 0.883-1.462 | .319 |
| Employment (compared to full time) | |||
| Disabled | 1.791 | 1.521-2.111 | <.001 |
| Part time | 1.230 | 1.059-1.429 | .007 |
| Retired | 1.039 | 0.933-1.156 | .490 |
| Student | 1.052 | 0.913-1.213 | .483 |
| Not employed | 1.396 | 1.279-1.522 | <.001 |
| Self employed | 1.050 | 0.856-1.287 | .642 |
| Marital status (compared to married) | |||
| Divorced or separated | 1.267 | 1.124-1.428 | <.001 |
| Significant other | 0.960 | 0.723-1.272 | .774 |
| Single | 1.206 | 1.114-1.306 | <.001 |
| Widowed | 1.178 | 1.007-1.378 | .041 |
| Tobacco use | 1.655 | 1.490-1.838 | <.001 |
| Alcohol use | 0.850 | 0.798-0.906 | <.001 |
| Drug use | 1.384 | 1.191-1.609 | <.001 |
| ZIP code population (compared to metro area, 1,000,000+) | |||
| Metro area w/250,000-1,000,000 | 0.762 | 0.629-0.925 | .006 |
| Metro area w/<250,000 | 0.608 | 0.430-0.859 | .005 |
| Urban, adjacent to metro area, >20,000 | 0.466 | 0.152-1.435 | .183 |
| Urban, not adjacent to metro area, >20,000 | 2.149 | 0.560-8.247 | .265 |
| Urban, 2500-19,999, adjacent to metro area | 0.876 | 0.572-1.342 | .543 |
| Urban, 2500-19,999, not adjacent to metro area | x | x | x |
| Completely rural, <2500, adjacent to metro area | 0.984 | 0.293-3.310 | .980 |
| Completely rural, <2500, not adjacent to metro area | 0.838 | 0.147-4.764 | .842 |
| ZIP code social vulnerability indices | |||
| Socioeconomic status theme | 0.781 | 0.599-1.019 | .069 |
| Household composition/disability theme | 3.394 | 2.066-5.575 | <.001 |
| Minority status and language theme | 1.040 | 0.755-1.431 | .812 |
| Housing type and transportation theme | 0.974 | 0.770-1.234 | .830 |
| New to Johns Hopkins dermatology | 1.427 | 1.321-1.541 | <.001 |
| Online patient portal status (active) | 0.671 | 0.579-0.777 | <.001 |
| Johns Hopkins primary care provider (yes) | 1.032 | 0.969-1.100 | .328 |
| Encounter day of week (compared to Monday) | |||
| Tuesday | 1.009 | 0.932-1.094 | .823 |
| Wednesday | 1.014 | 0.922-1.114 | .776 |
| Thursday | 0.953 | 0.879-1.034 | .247 |
| Friday | 0.915 | 0.836-1.002 | .055 |
| Encounter provider type (compared to physician) | |||
| Physician assistant | 1.233 | 1.067-1.425 | .005 |
| Resident | 1.027 | 0.908-1.162 | .668 |
AOR, Adjusted odds ratio.
Table II.
Demographics and characteristics of 99,476 encounters made in dermatology clinics at Johns Hopkins Medicine
| Variable | Show (N = 87,099) | Same-day cancel/no-show (N = 12,377) | P value |
|---|---|---|---|
| Age | <.001 | ||
| <18 | 1844 (2%) | 357 (3%) | |
| 18-30 | 12,151 (14%) | 2344 (19%) | |
| 31-45 | 18,695 (21%) | 2999 (24%) | |
| 46-60 | 21,656 (25%) | 3281 (27%) | |
| 61-75 | 24,233 (28%) | 2618 (21%) | |
| 76-90 | 7910 (9%) | 713 (6%) | |
| >90 | 608 (1%) | 67 (1%) | |
| Male sex | 33,714 (39%) | 4527 (37%) | <.001 |
| Employment status | <.001 | ||
| Disabled | 2716 (3%) | 984 (8%) | |
| Full-time | 38,877 (45%) | 4801 (39%) | |
| Part-time | 3223 (4%) | 519 (4%) | |
| Retired | 20,041 (23%) | 1897 (15%) | |
| Student | 4092 (5%) | 688 (6%) | |
| Not employed | 13,190 (15%) | 2838 (23%) | |
| Self-employed | 2070 (2%) | 229 (2%) | |
| Unknown | 2531 (3%) | 349 (3%) | |
| Race | <.001 | ||
| American Indian or Alaska Native | 192 (0%) | 43 (0%) | |
| Asian or Pacific Islander | 4400 (5%) | 496 (4%) | |
| White | 57,320 (66%) | 5438 (44%) | |
| Black or African American | 19,549 (6%) | 5502 (44%) | |
| Other | 5626 (6%) | 899 (7%) | |
| Hispanic or Latino Ethnicity | 3367 (4%) | 549 (5%) | .003 |
| Interpreter required | 1975 (2.3%) | 334 (2.7%) | .003 |
| ZIP code population (per county) | <.001 | ||
| Metro area w/1 million+ | 78,642 (90%) | 11,633 (94%) | |
| Metro area w/250,000-1,000,000 | 3214 (4%) | 305 (2%) | |
| Metro area w/<250,000 | 1014 (1%) | 85 (1%) | |
| Urban, 20,000+, adjacent to metro area | 82 (0%) | 6 (0%) | |
| Urban, 20,000+, not adjacent to metro area | 21 (0%) | 4 (0%) | |
| Urban, 2500-19,999, adjacent to metro area | 419 (0%) | 59 (0%) | |
| Urban, 1500-19,999, not adjacent to metro area | 48 (0%) | 0 (0%) | |
| Completely rural, <2500, adjacent to metro area | 42 (0%) | 5 (0%) | |
| Completely rural, <2500, not adjacent to metro area | 24 (0%) | 2 (0%) | |
| Unknown | 3591 (4%) | 280 (2%) | |
| Marital status | <.001 | ||
| Married | 45,330 (52%) | 4219 (34%) | |
| Divorced or separated | 6419 (7%) | 1221 (10%) | |
| Significant other | 816 (1%) | 102 (1%) | |
| Single | 27,915 (32%) | 5302 (43%) | |
| Widowed | 4321 (5%) | 566 (5%) | |
| Unknown | 2296 (3%) | 969 (8%) | |
| Tobacco use | 4898 (6%) | 2122 (17%) | <.001 |
| Alcohol use | 32,260 (37%) | 5018 (41%) | <.001 |
| Drug use | 1667 (2%) | 785 (6%) | <.001 |
On multivariable analysis, significant independent predictors of missed appointments were residence in ZIP codes with higher household composition and disability social vulnerability indices (AOR = 3.394; 95% CI, [2.066-5.575]), Black/African American race (AOR = 2.099; 95% CI, [1.945-2.265]), living on disability insurance (AOR = 1.791; 95% CI, [1.521-2.111]), new patient status (AOR = 1.427; 95% CI, [1.321-1.541]), and tobacco use (AOR = 1.655; 95% CI, [1.490-1.838]). Protective factors included online patient portal activation (AOR = 0.671; 95% CI, [0.579-0.777]), and living in a metro with 250,000 to 1,000,000 people (AOR = 0.762; 95% CI, [0.629-0.925]) and <250,000 people (AOR = 0.608; 95% CI, [0.430-0.859]). Notably, there was no significant difference in time between scheduling and appointment day, day of the week, provider type, or distance to clinic. Limitations include generalizability, sparse charted medical history, and lack of variables like weather or traffic, which may affect no-shows.
Missed appointments have been correlated with new patient status, race, smoking, and living in ZIP codes with poorer built environment or on disability.1,3 New patients may lack familiarity with providers or appointment logistics, leading to poorer attendance.3 Reasons racial minorities may not engage with health care may include medical mistrust or prior negative experiences; however, it is unclear why race alone would be predictive of making but not attending an appointment. Alongside increased medical comorbidities, smokers may also be more noncompliant with other aspects of their health, increasing their risk of no-shows.1 Disability and greater ZIP code household and disability social vulnerability indices are linked to increased costs, lost earning potential, and greater caretaker burden, potentially exacerbating barriers to appointment attendance. Residents in dense, large urban spaces may face more transportation constraints than those in smaller urban spaces, who have proximity to resources without crowding.4 Furthermore, online patient portal activation may be contingent on access to or familiarity with technology, decreasing no-show risk. These warrant further exploration of factors supporting appointment attendance.
Conflicts of interest
None disclosed.
Footnotes
Funding sources: Grant from Hopkins Business of Healthcare Initiative.
Preliminary abstract presented at Atlantic Derm Conference on April 23, 2022 (online).
IRB approval status: Reviewed and approved by the Johns Hopkins Medicine IRB (IRB00280674).
References
- 1.Dantas L.F., Fleck J.L., Cyrino Oliveira F.L., Hamacher S. No-shows in appointment scheduling – a systematic literature review. Health Policy. 2018;122(4):412–421. doi: 10.1016/j.healthpol.2018.02.002. [DOI] [PubMed] [Google Scholar]
- 2.Zou H., Hastie T. Regularization and variable selection via the elastic net. J R Stat Soc Series B Stat Methodol. 2005;67(2):301–320. doi: 10.1111/j.1467-9868.2005.00503.x. [DOI] [Google Scholar]
- 3.Pehr K. No show: incidence of nonattendance at a dermatology practice in a single universal payer model. J Cutan Med Surg. 2007;11(2):53–56. doi: 10.2310/7750.2007.00005. [DOI] [PubMed] [Google Scholar]
- 4.McLafferty S., Wang F. Rural reversal? Rural-urban disparities in late-stage cancer risk in Illinois. Cancer. 2009;115(12):2755–2764. doi: 10.1002/cncr.24306. [DOI] [PMC free article] [PubMed] [Google Scholar]