Secondary Logo

Journal Logo

Factors Associated With Missed Appointments at an Academic Pain Treatment Center

A Prospective Year-Long Longitudinal Study

Odonkor, Charles A., MD*; Christiansen, Sandy, MD; Chen, Yian, MD; Sathiyakumar, Asmitha, MD; Chaudhry, Hira, MD; Cinquegrana, Denise, MD; Lange, Jessica, MBA; He, Cathy, MD; Cohen, Steven P., MD†‡§‖

doi: 10.1213/ANE.0000000000001794
Healthcare Economics, Policy, and Organization: Original Clinical Research Report
Free

BACKGROUND: Interventional pain treatment centers represent an integral part of interdisciplinary care. Barriers to effective treatment include access to care and financial issues related to pain clinic operations. To address these challenges, specialty clinics have taken steps to identify and remedy missed clinic appointments. However, no prospective study has sought to identify factors associated with pain clinic “no-shows.”

METHODS: We performed a prospective, longitudinal year-long study in an inner-city, academic pain clinic in which patients scheduled for office visits and procedures were categorized as to whether they showed up or did not show up for their scheduled appointment without cancelling the day before. Twenty demographic (age, employment status), clinical (eg, diagnosis, duration of pain), and environmental (season, time and day of appointment) variables were assessed for their association with missing an appointment. The logistic regression model predicting no-shows was internally validated with crossvalidation and bootstrapping methods. A predictive nomogram was developed to display effect size of predictors for no-shows.

RESULTS: No-show data were collected on 5134 patients out of 5209 total appointments for a capture rate of 98.6%. The overall no-show rate was 24.6% and was higher in individuals who were young (<65 years), single, of ethnic minority background, received Medicare/Medicaid, had a primary diagnosis of low back pain or headaches, were seen on a day with rain or snow or for an initial consult, and had at least 1 previous pain provider. Model discrimination (area under curve) was 0.738 (99% confidence interval, 0.70–0.85). A minimum threshold of 350 points on the nomogram predicted greater than 55% risk of no-shows.

CONCLUSIONS: We found a high no-show rate, which was associated with predictable and unpredictable (eg, snow) factors. Steps to reduce the no-show rate are discussed. To maximize access to care, operation managers should consider a regression model that accounts for patient-level risk of predictable no-shows. Knowing the patient level, no-show rate can potentially help to optimize the schedule programming by staggering low- versus high-probability no-shows.

Published ahead of print March 8, 2017.

From the Departments of *Physical Medicine & Rehabilitation, Anesthesiology & Critical Care Medicine, and Neurology, Johns Hopkins School of Medicine, Baltimore, Maryland; and Departments of §Anesthesiology and Physical Medicine & Rehabilitation, Uniformed Services University of the Health Sciences, Bethesda, Maryland.

Published ahead of print March 8, 2017.

Accepted for publication November 2, 2016.

Funding: This work was funded in part by the Centers for Rehabilitation Sciences Research, Bethesda, MD; grant number HU0001-15-2-0003. The role of the funding sources was to provide support for research personnel.

The authors declare no conflicts of interest.

The opinions or assertions contained herein are the private views of the authors and are not to be construed as official or as reflecting the views of the Department of the Army or the Department of Defense.

Reprints will not be available from the authors.

Address correspondence to Steven P. Cohen, MD, Department of Anesthesiology & Critical Care Medicine, Uniformed Services University of the Health Sciences, 550 North Broadway, Suite 301, Bethesda, MD. Address e-mail to scohen40@jhmi.edu.

Pain represents a major public health burden, costing the United States an estimated $560–$635 billion annually in health care costs and economic productivity.1 Approximately 100 million Americans experience chronic pain with low back pain constituting the leading cause of years lost to disability.1,2 Research has shown that low socioeconomic status (SES) is associated with an increased prevalence rate of pain.3–6 Increasing access of underserved patients with chronic pain to pain management clinics may represent an important step to address this societal burden; however, large-scale cohort studies have found that patients of low SES are more likely to miss scheduled appointments than middle- or upper-class cohorts.7

Missed and late appointments constitute a significant loss of efficiency and resources in health care. Between 2012 and 2013, the British National Health Service reported a missed appointment rate of approximately 9.3%, costing hundreds of millions of pounds.8 Many studies have examined variables contributing to missed appointments in both primary care and subspecialty clinics. These studies have shown factors such as lower SES, race, younger age, being an immigrant, caring for a sick relative/lack of child care, health care disparities, transportation difficulties, being unemployed, and receiving Medicaid to be associated with a lower attendance rate.9–14 Identification of such demographic, clinical, and environmental characteristics can facilitate targeted interventions that can potentially help reduce missed and late appointments and subsequently improve treatment outcomes and conserve resources.

Studies focusing on missed appointment characteristics among the chronic pain population are limited. A recent literature review yielded a single retrospective study by Shaparin et al,15 which revealed that characteristics such as Spanish as a primary language and increased geographic distance from the clinic were associated with a greater likelihood of missed appointments, whereas patients scheduling appointments for a particular complaint versus a nonspecific complaint, appointments for a procedure, being unemployed, and greater continuity of care as measured by the number of previous visits were associated with an increased chance of attending an appointment at an urban academic pain clinic. Although this study provides an overview of missed appointment characteristics in a chronic pain population, it is limited by its retrospective nature and the observation that a majority of patients were either on Medicaid or received hospital charity care. Therefore, observations from the study by Shaparin et al15 are likely relevant to an underserved population but may not be generalizable to other pain clinic settings catering to more diverse populations.

Studies examining no-show rates in pain clinics may yield different findings than those conducted in other settings because patients with chronic pain may be more likely to be on medical assistance or disability, have multiple comorbidities, and contend with a host of psychosocial and socioeconomic issues that can affect reliability and punctuality.4,5,15–17 Moreover, there is a dearth of predictive risk tools that analyze the likelihood of no-shows for patients presenting to pain clinics. Because single factors rarely predict prognosis accurately, tools such as nomograms have been proposed as a practical guide for more precise prognostication in other settings.18

To address these limitations, we performed a prospective study assessing show rates for all pain clinic appointments over a 1-year period at Johns Hopkins Hospital in Baltimore, Maryland, a large, inner-city, academic pain clinic that nevertheless caters to a diverse population. We examined a host of demographic, clinical, socioeconomic, and environmental factors to include residential distance from the clinic, employment status, weather status at the time of the appointment, primary insurance type, appointment type, time of day and year, holiday occurrence, primary diagnosis, opioid use, and average pain scores as possible contributors to missed appointments. Our hypotheses were that the missed appointment rate in our inner-city, academic pain clinic would exceed those in other settings and patient populations and that we would be able to identify certain factors that predisposed patients to miss their appointment. By elucidating the factors associated with nonattendance, and the barriers patients face in attending scheduled appointments, we hope that practitioners and administrative staff can optimize the availability of pain management resources for patients with chronic pain.

Back to Top | Article Outline

METHODS

This study was designed as a quality improvement project in conjunction with the business development team of the Department of Anesthesiology & Critical Care Medicine and approved by the Johns Hopkins institutional review board.

Back to Top | Article Outline

Primary Data Collection

Primary data for this prospective, observational study were collected between January 1, 2015, and December 31, 2015. The questionnaire form was designed to include a multitude of demographic, clinical, and environmental factors that were hypothesized to affect the “no-show” rate at the Johns Hopkins Blaustein Pain Treatment Center. These factors included age, gender, marital status and annotated race of the patient, duration and intensity of pain, primary pain complaint, whether the patient was receiving opioid therapy at the time of the appointment, whether he or she had previously seen another pain provider for the same condition, season, time of day (afternoon or morning), day of the week of the appointment, how far in advance the appointment was scheduled, whether the appointment was scheduled in proximity to a holiday, precipitation during time of the appointment, employment status, distance of primary residence from the pain treatment center, primary insurance type, and type of appointment. The questionnaires were printed and inserted into all scheduled patient charts before their appointment time. Providers, nursing staff, and front office staff all assisted patients in completing the form during their clinic or procedure appointment. Data were collected on all scheduled appointments. When indicated (ie, for “no-shows,” appointment scheduling, etc), physicians and research staff collected data by reviewing electronic medical records (Epic Systems, Verona, WI) and consult requests and directly questioning the patient in person or via a telephone call.

Back to Top | Article Outline

Definitions

Patients were classified as “no-shows” when they either did not show up for an office visit or procedure or cancelled the appointment on the day of the appointment. Types of appointments evaluated included new patient consults, follow-up office visits, and procedure appointments. Follow-up visits were further classified based on whether the appointment was scheduled with a physician or physician’s assistant.

The relationship between how far in advance an appointment was scheduled and the outcome was evaluated by categorizing the variable “appointment scheduled” into late (<1 week), intermediate (1–4 weeks), and early (>4 weeks before the scheduled appointment). A visit was considered to be a “holiday appointment” when it occurred on the workday before or after an observed federal holiday (eg, the Friday and Tuesday preceding and after Memorial and Labor Day) or on the date of an unobserved federal holiday (Columbus and Veterans Days). Insurance type was classified as being Medicare, a state Medicaid carrier, private insurance, military (Tricare), or self-pay. Primary residence was divided into “in-state” or “out-of-state.”

Back to Top | Article Outline

Data Coding

After all information was verified; data were coded onto an Excel spreadsheet (Microsoft, Redmond, WA) and deidentified. Numerical codes were used to reduce information to categorical variables and dummy coding for multilevel categories. For practical purposes and ease of interpretation of predictive models, selected variables were dichotomized as outlined: age (<65 vs ≥65 years), holiday week (yes versus no), time of day for appointment (morning versus afternoon), and day of the week (Friday versus non-Friday).

Back to Top | Article Outline

Statistical Analysis

Data were summarized with descriptive statistics. Group means (show vs no-show) for continuous and categorical variables were compared via Student t tests and χ2 and Fisher exact tests, respectively. Similar analyses were performed for subjects with missing data by comparing the group means and distributions of predictor variables for subjects missing data versus those of subjects with complete data. Correlations among independent variables and covariates were evaluated by Spearman rank correlation coefficient. For any 2 significant collinear variables (r ≥ 0.3), the 1 contributing most to variance for no-show rate was chosen for inclusion in the final prediction models. Hypotheses-based models were generated using previously defined explanatory variables (age, gender, ethnicity, marital status, employment status, and insurance) and covariates identified by univariate analysis (precipitation, season, opioid usage, diagnosis, and timing and type of appointment). Multivariable logistic regression models were developed in a manual backward elimination fashion (PROC Logistic; SAS Statistical package version 9.3®, Cary, NC) to identify factors associated with no-shows.19 Interaction effects were tested via a set of separate interaction–inclusive models to determine associations with the primary outcome. Probability of no-show was determined based on logistic regression model β estimates by the equation:

where βk represents beta and xk is set of predictor variables

Model prediction of no-shows was internally validated using crossvalidation and bootstrapping methods with 2 × 104 iterations. The area under the curve was used for model discrimination.

Factors associated with no-shows based on the final main effects multivariable logistic regression model were used to develop a prognostic nomogram. Predictors were assigned points based on the estimated regression coefficients. Disregarding statistical significance and direction (absolute β values), the predictor with the largest impact (β value) in the model was assigned the highest point (on a 0–100 scale). All other predictors were scored in a sequential fashion in proportion to the points assigned to the predictor with the largest impact (β value). As an example, the predictor with the largest β, headache (β = 1.7), was assigned 100 points and cancer pain (β = 1.02) was assigned 60 points ([1.02/1.7] × 100).

Once all factors from the model had been assigned points and ranked via a linear prediction method, based on estimated β coefficients, total points were obtained and a linear projection was performed to obtain the probability scale (0–1).

A nomogram was generated from the points per variable as well as total points and the predicted probability scale. The goodness-of-fit of the nomogram was calculated using the Hosmer–Lemeshow test. No significant interaction terms were identified. Data were found to be missing at random (nonresponse observed for unemployed and older patients) and completely at random (probability of missing data was the same for all units). Missing data were addressed using routine multivariate imputation methods to fit candidate multivariate models with imputed missing data. All statistical analyses were performed via SAS software, version 9.3 (Cary, NC) using a 2-sided hypothesis test with probability of a type I error set at 0.01 (Bonferroni correction for multiple comparisons).

Back to Top | Article Outline

RESULTS

Baseline Characteristics

Table 1

Table 1

Figure 1

Figure 1

Of 5209 total appointments over the 2015 calendar year, 5162 were approached for survey participation. Data were collected on 5134 patients for a capture rate of 98.6%. There were 1262 no-shows of which 940 did not call to cancel and 322 called to cancel within 24 hours before the scheduled appointment (Figure 1). For all collected surveys, 469 charts contained incorrect or missing data as a result of nonresponse or incorrectly documented information for a missing rate of less than 10% across tested variables (Table 1). Characteristics of participants with missing values compared with those of participants with complete data revealed that the 2 groups were not significantly different in age, gender, marital status, duration of chronic pain, average pain score, time of appointment (morning versus afternoon), diagnosis, presence of prior pain provider(s), profession (white collar, blue collar, or disability), and use of opioids. Among the most commonly treated pain categories, low back pain (n = 1963), neck pain (n = 739), and joint pain (n = 559) were the most frequently encountered conditions.

Back to Top | Article Outline

Factors Associated With No-Show Rate

Table 1 presents demographic, environmental, and clinical characteristics of the study population stratified by whether or not they showed up for their appointments. The group of subjects who failed to show up for their appointments was comprised mostly of women (56.5%), minorities (55.2% black), individuals who were unemployed (70.3%), people with low back pain (40.2%), and patients residing within the state of Maryland (66.3%). On average, most subjects who did not show up for their appointments were younger (50.0 ± 13.0 years) compared with those who did attend their visit (54.0 ± 15.0 years, P < .0001), used opioids for pain management (51.6%, P = .02), were on disability or received Workers’ Compensation (36.5%, P < .0001), and were receiving Medicare (39.7%) or Medicaid (28.7%) as their primary insurance (P < .0001). Among 1963 subjects with low back pain, 507 failed to show up for their appointments, a no-show frequency of 25.8% (P = .001). Other differences between the 2 groups are highlighted in Table 1. The overall no-show rate for all appointments was 24.6%.

Back to Top | Article Outline

Reasons Cited for Not Showing Up for Appointment

Whenever possible (n = 469), we tried to obtain the reason why patients missed appointments from a set group of answers. The top 5 reasons for no-shows were patient overslept (34.7%), lack of transportation (14.4%), personal illness (13.6%), work/school conflicts (12.8%), and time conflicts/patient mistakenly double-booked another appointment (12.8%).

Back to Top | Article Outline

Multivariable Logistic Regression

Table 2

Table 2

Table 2 shows the results of the multivariable logistic regression model predicting the likelihood of clinical no-shows. The model was adjusted for the following variables: SES including age, gender, ethnicity, health care payer, pain diagnosis, number of pain providers, time-related variables (season and precipitation), and the type of appointment. The P value of the Hosmer–Lemeshow test for the prediction model was .15, which indicated good statistical fit of the model. Although factors such as employment status (odds ratio [OR] = 1.2, 99% confidence interval [CI], 1.03–4.6), opioid use (OR = 1.18, 99% CI, 1.02–4.4), and time of appointment (OR = 1.2, 99% CI, 1.05–6.3) were significant factors in univariate analysis, they were excluded from the final model as a result of high multicollinearity with ethnicity and insurance (Spearman r >0.35). Separate models where these 3 variables replaced the 2 SES measures showed that opioids (P = .09), employment status (P = .046), and the time of appointment (P = .10) did not achieve statistical significance.

Back to Top | Article Outline

Prediction Nomogram for No-Shows

Table 3

Table 3

Figure 2

Figure 2

Figure 2 highlights the results of the prognostic nomogram derived from the multivariable logistic regression model. Ten key predictors were identified to predict no-shows within 74% accuracy (area under the curve for model = 0.738, 99% CI, 0.72–0.80). At a cutoff threshold of 0.55% probability of no-show, the sensitivity and specificity of nomogram prediction were 20% and 96.5%, respectively. The minimal threshold score needed to yield greater than 55% prediction risk of no-shows was 350 points. Table 3 illustrates a practical example of the nomogram in predicting risk of no-show for 2 patients, A and B, who differ by diagnosis, marital status, and type of appointment. Mr. A is a 40-year-old, single Latino patient with persistent chronic migraines, who is scheduled for a consult. Mr. B is a 70-year-old married man on Medicare who presents for follow-up for management of chronic low back pain. Using the nomogram, Mr. A and B score 342 points and 196 points, respectively. Because Mr. A has a 1.5 times higher risk of no-show than Mr. B, a clinician may consider staggering the appointment given to Mr. A with that of a second patient with a lower predicted risk of no-show.

Back to Top | Article Outline

DISCUSSION

In this longitudinal study, we observed a 24.6% no-show rate over a 1-year period. Similar to our hypotheses, a number of factors were shown to be associated with an increased no-show rate, including younger age, ethnic minority background, marital status, health payer, diagnosis, number of previous pain providers, season and precipitation status on the day of appointment, and the type of appointment.

Back to Top | Article Outline

Comparison to Other Studies

Our no-show study rate is higher than those reported in a variety of different specialty clinics, in children and adults, within and outside of the United States,7,10,20,21 yet many of the major findings mirrored those observed in similar studies performed in other settings. In a similarly sized (n = 5608) retrospective study evaluating missed appointments in a community health center serving a low-income population with a high percentage of immigrants, Kaplan-Lewis and Percac-Lima10 found that minorities, those with lower incomes, and those receiving Medicaid assistance were more likely to miss appointments. In this study, forgetting and miscommunication were the 2 most commonly cited reasons for not showing up. We did not find a significant difference in no-show rate based on gender. In studies examining factors associated with missed appointments, some have shown men,14 while others have shown women13 to be more likely to not show up for doctor visits. Lehmann et al11 evaluated factors associated with missed appointments in a university-based internal medicine clinic in Switzerland over a 1-month period. Among the 1296 scheduled appointments, the no-show rate was 15.8%. Similar to our findings, variables associated with missed appointments included being an immigrant (non-European) and younger age. However, their results differed from ours in that follow-up visits were associated with a higher rate of missed appointments than consults. Not surprisingly, our results were also similar to a retrospective study performed over a 9-year period in a pediatric autism clinic at an inner-city affiliate of Johns Hopkins Medical Institutions.22Among 8049 children scheduled for 43,504 appointments, the authors found that black race, receiving medical assistance, and a longer lead/wait time were associated with increased rates of no-show or cancellation (increased lead time), respectively. Longer lead times have been shown in other clinic types to increase no-show rate, possibly as a result of spontaneous improvement, increased forgetfulness, or the patient seeking alternative medical care.23 Concordant with the results of Lehmann et al,11 no-show rates were higher for follow-up than for initial evaluations (15% vs 9%). Our results should also be viewed in comparison with those of Shaparin et al,15 who performed a retrospective study evaluating 3408 patient encounters over a 4-year period at a pain clinic located in inner city Newark, New Jersey. They reported a 48% rate of missed appointments with only 26% of those cancelling in advance. Shaparin et al15 found that speaking Spanish as the primary language and living in close proximity to the clinic increased the likelihood of not showing for an appointment, whereas having a procedure or follow-up appointment and being unemployed was associated with a lower risk for missing an appointment. In contrast, our study indicates that although being unemployed alone increased the risk for missing an appointment, when adjusted for other SES covariates, unemployment status was not a significant predictor of no-show rates. The retrospective nature and relatively small sample size of the study by Shaparin et al15 over a 4-year period, which likely contained a high proportion of missing data, raises questions about the generalizability of their study results.

As health care transitions toward value-based purchasing and delivery, there is increasing utility of prospective predictive risk modeling in testing and designing strategies to improve health care quality and outcomes. Our study provides a practical tool to help guide clinicians determine the most favorable iteration of factors with the lowest predicted probability of no-shows. Previous studies suggest that patients prefer “advanced access” and shorter lead time appointment scheduling as reflected by lower no-show rates.23,24 Although others have proposed various scheduling schema for clinic appointments,25–27 our study provides an internally validated model for individualized no-show predictive risk modeling for an interventional pain clinic.

Back to Top | Article Outline

Explanation of Findings

Several of our findings warrant further discussion. A consistent finding among similar studies suggests that racial and ethnic minorities have higher no-show rates than nonminorities,10–12,21 which may reflect both socioeconomic factors that can adversely affect health behavior and treatment adherence as well as issues related to trust and poor rapport in an environment in which most doctors do not share the same physical and cultural characteristics as their patients. For example, disadvantaged patients of any racial or ethnic background (eg, white people from Appalachia), who face myriad logistic (eg, transportation, childcare), economic (eg, copays), professional (eg, work conflicts), and social (eg, escort requirement) barriers, are likely to have a higher rate of missed appointments. A study performed at a Johns Hopkins HIV clinic found that black patients reported lower levels of trust in their providers, resulting in decreased compliance.28 Studies examining reasons for this distrust have found that nonconcordance between patients and providers in ethnic and racial characteristics are associated with lower levels of patient trust and satisfaction.29

In a finding that might appear counterintuitive, individuals who were employed, particularly those in white-collar professions, were more likely to show up for appointments. This is probably the result of the fact that these individuals are less likely to have to deal with other factors such as loss of insurance, transportation issues, or taking care of a sick relative that might increase the risk of missing an appointment. Although not statistically significant and excluded from the proposed final model as a result of colinearity with racial, insurance, and employment effects, univariate analysis showed that there was a trend toward an association between out-of-state residence and a lower likelihood of missing an appointment (P = .08), which is in contradistinction to the findings of Shaparin et al.15 This may be attributable to the fact that most individuals treated at Johns Hopkins from out-of-state come specifically to see a particular doctor or were referred specifically to the Blaustein Pain Treatment Center and therefore may have higher expectations. However, out-of-state residency was not a robust predictor.

Back to Top | Article Outline

Steps to Reduce the No-Show Rate and Maximize Operations

There are several steps that can be taken to reduce the no-show rate and maximize the attendance including: telephone and mobile messaging reminders sent out between 1 and 3 days before the appointment, which may result in higher show rates than postal (or no) reminders,30–32 decreasing the interval time between when the appointment is made and when the visit occurs,23 hiring more diverse staff members with geographic or cultural connections to the patient population, and improving patient satisfaction (eg, improving physician trust and reducing physician wait times), which may reduce no-show rates for follow-up visits and procedures.28 Similar to the model used by airlines and admission committees, using a logistic regression model that takes into account predicted no-shows and overbooks was shown in 1 study to significantly improve efficiency.33

Back to Top | Article Outline

Limitations

There are several limitations to our study that should be noted. First, although this was a prospective study in which our overall capture rate exceeded 98%, there were missing data points for specific variables, including marital status, location of residence, employment status, profession, insurance, and opioid use. Therefore, the results of these findings, particularly those not shown to be associated with no-show rate, may have underestimated the effects of these factors. Second, the Blaustein Pain Treatment Center is located in an inner city with a 68% minority population, treats a high proportion of patients with intractable pain who failed other treatments, and may not be representative of most other private practice and academic pain centers. Third, our prognostic nomogram needs to be externally validated. Nonetheless, our findings provide a helpful tool, which may potentially guide clinicians and administrations making appointment decisions in pain clinic settings.

Back to Top | Article Outline

CONCLUSIONS

In summary, in the first prospective study evaluating pain clinic no-show rates from an academic teaching hospital, we found a rate of 24.6%, which is higher than reported rates in nonpain clinic settings. The factors we found to be associated with a higher no-show rate (ie, minority and lower SES) have been shown in other studies to predict missed appointments; however, these findings also highlight new factors such as having more than 1 previous pain provider, and provide no-show risks for common presenting pain diagnoses. Simple steps to reduce the no-show rate may include sending out electronic messages within 72 hours of the appointment and staggered appointments. To maximize utilization, clinic managers should consider a model that takes into account the individual-level risk of a patient not showing up for their appointment.

Back to Top | Article Outline

DISCLOSURES

Name: Charles A. Odonkor, MD.

Contribution: This author collected data, prepared the manuscript, and reviewed the manuscript. This author helped with statistical analysis.

Name: Sandy Christiansen, MD.

Contribution: This author collected data, prepared the manuscript, and reviewed the manuscript.

Name: Yian Chen, MD.

Contribution: This author collected data, prepared the manuscript, and reviewed the manuscript.

Name: Asmitha Sathiyakumar, MD.

Contribution: This author collected data, prepared the manuscript, and reviewed the manuscript.

Name: Hira Chaudhry, MD.

Contribution: This author collected data and reviewed the manuscript.

Name: Denise Cinquegrana, MD.

Contribution: This author collected data and reviewed the manuscript.

Name: Jessica Lange, MBA.

Contribution: This author collected data and reviewed the manuscript.

Name: Cathy He, MD.

Contribution: This author collected data and reviewed the manuscript.

Name: Steven P. Cohen, MD.

Contribution: This author collected data, prepared the manuscript, and reviewed the manuscript. This author helped with statistical analysis and study design.

This manuscript was handled by: Nancy Borkowski, DBA, CPA, FACHE, FHFMA.

Back to Top | Article Outline

REFERENCES

1. Institute of Medicine (US) Committee on Advancing Pain Research, Care, and Education. Relieving Pain in America: A Blueprint for Transforming Prevention, Care, Education and Research. 2011.Washington, DC: The National Academies Press.
2. Murray CJ, Atkinson C, Bhalla K, et alUS Burden of Disease Collaborators. The state of US health, 1990–2010: burden of diseases, injuries, and risk factors. JAMA. 2013;310:591–608.
3. Aggarwal VR, Macfarlane TV, Macfarlane GJ. Why is pain more common amongst people living in areas of low socio-economic status? A population-based cross-sectional study. Br Dent J. 2003;194:383–387.
4. Andersson HI, Ejlertsson G, Leden I, Rosenberg C. Chronic pain in a geographically defined general population: studies of differences in age, gender, social class, and pain localization. Clin J Pain. 1993;9:174–182.
5. Papageorgiou AC, Macfarlane GJ, Thomas E, Croft PR, Jayson MI, Silman AJ. Psychosocial factors in the workplace—do they predict new episodes of low back pain? Evidence from the South Manchester Back Pain Study. Spine (Phila Pa 1976). 1997;22:1137–1142.
6. Locker D. The burden of oral disorders in a population of older adults. Community Dent Health. 1992;9:109–124.
7. Majeroni BA, Cowan T, Osborne J, Graham RP. Missed appointments and Medicaid managed care. Arch Fam Med. 1996;5:507–511.
8. Hallsworth M, Berry D, Sanders M, et alStating appointment costs in SMS reminders reduces missed hospital appointments: findings from two randomised controlled trials. PLoS One. 2015;10:e0137306.
9. Campbell JD, Chez RA, Queen T, Barcelo A, Patron E. The no-show rate in a high-risk obstetric clinic. J Womens Health Gend Based Med. 2000;9:891–895.
10. Kaplan-Lewis E, Percac-Lima S. No-show to primary care appointments: why patients do not come. J Prim Care Community Health. 2013;4:251–255.
11. Lehmann TN, Aebi A, Lehmann D, Balandraux Olivet M, Stalder H. Missed appointments at a Swiss university outpatient clinic. Public Health. 2007;121:790–799.
12. Brown KA, Shetty V, Delrahim S, Belin T, Leathers R. Correlates of missed appointments in orofacial injury patients. Oral Surg Oral Med Oral Pathol Oral Radiol Endod. 1999;87:405–410.
13. Waller J, Hodgkin P. Defaulters in general practice: who are they and what can be done about them? Fam Pract. 2000; 17:252–253.
14. Simmons AV, Atkinson K, Atkinson P, Crosse B. Failure of patients to attend a medical outpatient clinic. J R Coll Physicians Lond. 1997;31:70–73.
15. Shaparin N, White R, Andreae M, Hall C, Kaufman A. A longitudinal linear model of patient characteristics to predict failure to attend an inner-city chronic pain clinic. J Pain. 2014;15:704–711.
16. Green CR, Anderson KO, Baker TA, et alThe unequal burden of pain: confronting racial and ethnic disparities in pain. Pain Med. 2003;4:277–294.
17. Nguyen M, Ugarte C, Fuller I, Haas G, Portenoy RK. Access to care for chronic pain: racial and ethnic differences. J Pain. 2005;6:301–314.
18. Iasono A, Schrag D, Raj GV, Panageas K. How to build and interpret a nomogram for cancer prognosis. J Clin Oncol. 2008;26:1364–1370.
19. Patterson ES, Doebbeling BN, Fung CH, Militello L, Anders S, Asch SM. Identifying barriers to the effective use of clinical reminders: bootstrapping multiple methods. J Biomed Inform. 2005;38:189–199.
20. Jeppesen MH, Ainsworth MA. Telephone reminders reduced the non-attendance rate in a gastroenterology outpatient clinic. Dan Med J. 2015;62:A5083.
21. Markowitz JT, Volkening LK, Laffel LM. Care utilization in a pediatric diabetes clinic: cancellations, parental attendance, and mental health appointments. J Pediatr. 2014;164:1384–1389.
22. Kalb LG, Freedman B, Foster C, Menon D, Landa R, Kishfy L, Law P. Determinants of appointment absenteeism at an outpatient pediatric autism clinic. J Dev Behav Pediatr. 2012;33:685–697.
23. McMullen MJ, Netland PA. Lead time for appointment and the no-show rate in an ophthalmology clinic. Clin Ophthalmol. 2015;18:513–516.
24. Rose KD, Ross JS, Horwitz LI. Advanced access scheduling outcomes: a systematic review. Arch Intern Med. 2011;171:1150–1159.
25. Huang Y, Hanauer DA. Patient no-show predictive model development using multiple data sources for an effective overbooking approach. Appl Clin Inform. 2014;5:836–860.
26. Mehrotra A, Keehl-Markowitz L, Ayanian JZ. Implementing open-access scheduling of visits in primary care practices: a cautionary tale. Ann Intern Med. 2008;148:915–922.
27. Patrick J. A Markov decision model for determining optimal outpatient scheduling. Health Care Manag Sci. 2012;15:91–102.
28. Saha S, Jacobs EA, Moore RD, Beach MC. Trust in physicians and racial disparities in HIV care. AIDS Patient Care STDS. 2010;24:415–420.
29. Street RL Jr, O’Malley KJ, Cooper LA, Haidet P. Understanding concordance in patient-physician relationships: personal and ethnic dimensions of shared identity. Ann Fam Med. 2008;6:198–205.
30. Gurol-Urganci I, de Jongh T, Vodopivec-Jamsek V, Atun R, Car J. Mobile phone messaging reminders for attendance at healthcare appointments. Cochrane Database Syst Rev. 2013;12:CD007458.
31. Rohman L, Maruswezki D, Boyce Cam N. The impact of a text messaging service on orthopaedic clinic did not attend rates. J Telemed Telecare. 2015;21:408–413.
32. Leong KC, Chen WS, Leong KW, et alThe use of text messaging to improve attendance in primary care: a randomized controlled trial. Fam Pract. 2006;23:699–705.
33. Daggy J, Lawley M, Willis D, et alUsing no-show modeling to improve clinic performance. Health Informatics J. 2010;16:246–259.
© 2017 International Anesthesia Research Society