Effect of Telephone Call and Text Message Reminders on Patient Return to Acupuncture Follow-Up Treatment: A Pilot Randomized Controlled Trial

Abstract

Background: Telephone calls and text messages function as cues to elicit patient behavior.

Objective: We tested the effect of telephone call and text message reminders on patient return to acupuncture follow-up treatment.

Design: This is a randomized controlled trial.

Setting and Subjects: We recruited adults visiting an acupuncture clinic for a new treatment consultation. Our sample contained 120 participants with 40 per study group.

Interventions: Consenting patients were randomized to 1 of 3 study exposures: single voice call, single text message, or treatment as usual (TAU/no reminder). Exposures were sent 3 days after patient's initial treatment.

Outcome Measures: Data from clinic charts were abstracted to quantify the absence/presence of a follow-up treatment return in the 30 days after initial treatment. Participants provided self-report of pain symptoms 10 and 30 days after initial treatment on the pain disability index (PDI) to measure change in PDI by return to follow-up treatment as a secondary outcome.

Results: Telephone call (56%, P = 0.98) and text message (57%, P = 0.99) groups showed similar proportion of follow-up treatment returns compared with TAU group (57%). Presence of a follow-up appointment scheduled at the initial treatment predicted patient treatment return (odds ratio: 5.87, P < 0.01). Follow-up treatment return predicted reduced PDI scores at day 30 (β = −3.09, P = 0.02).

Conclusions: Adding a 1-time telephone call or text message reminder to standard clinic practice did not improve patient return to acupuncture follow-up treatment within 30 days of initial treatment. Scheduling a future appointment date at initial treatment visit may improve treatment return, and return attendance appears protective of pain disability in the short term.

Introduction

Annually, 3.4 million U.S. adults receive acupuncture treatment.¹ Acupuncture patients are often recommended by their clinician to return for follow-up treatment to manage their clinical symptoms. This is reasonable advice given that follow-up treatment return is associated with improvement in pain severity and reduction in disability.^2,3 For conditions such as musculoskeletal pain, follow-up treatment return is important given musculoskeletal conditions require multiple treatment sessions for patients to experience therapeutic benefits over time.^4,5 However, follow-up treatment return is modest in acupuncture clinics. Only 50% of patients return for a follow-up visit⁶ or complete their recommended treatment sessions.⁷ Low rates of treatment return can limit the health benefits patients may receive from acupuncture.

Despite the potential benefits of completing an acupuncture treatment regimen, only a handful of studies have examined factors that predict follow-up treatment return. Bishop⁷ found that patients who report more severe symptoms and require considerable effort to attend a treatment session have lower treatment completion rates. Factors attributed to noncompletion of treatment include worsening of symptoms, perceiving treatment as ineffective, losing coverage, and not expecting additional benefits from treatment.^8–11 Of the factors that predicted follow-up treatment return, receiving a treatment plan at the first clinic visit for a new consultation resulted in more follow-up visits.⁶ Beyond these preliminary findings, there is no study we are aware of that tests reminders on patient return to acupuncture treatment.

Phone-based reminders are widely used to improve patient attendance to medical appointments. These reminders serve as cues to action¹² and are used to encourage attendance to follow-up appointments, improve medication adherence, and support recommended lifestyle changes.^13–15 Two mobile reminders, telephone call and text message, can improve attendance to health care appointments.^16,17 Arora showed that the proportion of patients who returned to a follow-up visit was higher in those who received text message reminders than those in a no reminder control (n = 328, 73% vs. 62%, P = 0.05).¹⁸ No study has yet to use an experiment to test whether mobile phone reminders improve patient return to acupuncture treatment.

This pilot study aims to determine whether mobile phone reminders improve patient return to acupuncture follow-up treatment. The study tests whether a single telephone call or text message reminder increases patient return to follow-up treatment within 30 days of initial treatment consultation. We hypothesize that the proportion of follow-up treatment returns will be higher in telephone call and text message groups than in no reminders. As an exploratory analysis, we test, among those presenting with a pain conditions at baseline, the effect of follow-up treatment return on pain scores during a 30-day period. We hypothesize that treatment return will predict reduction in pain scores.

Methods

Trial Design

We conducted a single-site 3-group randomized controlled trial (RCT). Participants were randomized with age as a strata variable (<50 and ≥50 years) in 1:1:1 allocation ratio (telephone call: text message: treatment as usual [TAU]) on the date of study enrollment (same day as the initial consultation) using a computerized random sequence generator. The age strata was used to account for a potentially higher number of treatment returns among older patients.⁶ All participants provided written informed consent. The a priori sample size was set to 40 for each group. The University of Southern California Institutional Review Board (HS-17-00008) approved all study procedures (Trial registration: clinicaltrial.gov, Identifier: NCT03645083).

Participants

The study was conducted in Los Angeles, California, from February to September 2017. Adult patients seeking treatment at an acupuncture teaching clinic were recruited by a research assistant at the front desk upon clinic arrival. Inclusion criteria were age 18 years or older, seeking a new treatment consultation, owner of a mobile phone with text message capability, and receiving treatment at the intern clinic that uses a standardized treatment protocol (i.e., trainee supervised by a licensed acupuncturist). Clinic staff and trainees were excluded from participating in the study.

Procedures

Patients were enrolled in the study on the day of their initial treatment consultation (day 0). Upon clinic arrival, patients interested in the study were screened to determine eligibility and then enrolled. After the treatment session, all participants completed a self-report baseline questionnaire. Participants were assigned to study groups on the evening of day 0. Three days later (day 3), participants received their assigned study exposure and so became aware of their study group. To assess the primary outcome, our study team assessors conducted chart reviews 30 days after the initial treatment (day 30) to quantify the absence or presence of a follow-up treatment return during this 30-day follow-up period. They were instructed to transcribe data exactly as reported in the medical chart.

In addition to the primary RCT protocol, the study research assistant sent all participants an e-mail with a link to complete an online follow-up questionnaire about their treatment progress 10 and 30 days after their initial treatment (days 10 and 30). Survey data from days 0, 10, and 30 allowed for repeated measures analyses of 2 measures, the visual analog scale (VAS) and pain disability index (PDI). All participants were compensated with a $5 gift card for completing the baseline questionnaire. Those who completed 1 of the follow-up questionnaires were then eligible to win a voucher for an acupuncture treatment visit after the study was completed.

Interventions

Telephone Call Reminder Group

On day 3 of the study, a research assistant made up to 3 telephone calls to reach the participants. When connected, the research assistant recited a scripted greeting and the intervention message. The message stated: “We encourage all patients to schedule a follow-up visit to help with their symptom management.” If the call was not connected after 3 attempts, in the final attempt, the research assistant left a voicemail with the same intervention message. Receipt of the exposure was confirmed when the participant answered the call.

Text Message Reminder Group

On day 3 of the study, a research assistant sent up to 3 text messages to participant's preferred phone number. The text message stated: “We encourage all patients to schedule a follow-up visit to help with their symptom management.” Receipt of the exposure was confirmed when the participant text-replied as prompted.

TAU Group

The TAU group followed the typical clinical routine practice at the study site. No reminder was sent out to patients at this clinic regarding their existing treatment or follow-up appointment during the study period. As such, the TAU group did not receive a reminder message in any form.

Measures

Follow-Up Treatment Return

The primary study outcome was patient return to a follow-up treatment visit within 30 days of initial treatment. Return is defined as the participant returned to the clinic to continue acupuncture on the same chief complaint as the initial treatment visit. The research assistant reviewed clinic charts to identify participants who had the absence or presence of records for a follow-up treatment visit within 30 days after their initial treatment consultation. All 120 participants enrolled in the study had these data available. The authors generated 2 outcome variables for follow-up treatment return: (1) a binary outcome coded 1 for ≥1 follow-up treatment visits and 0 for no return visit, and (2) total count of follow-up visits within the 30-day period after the initial treatment.

Participant Characteristics

Baseline demographic factors assessed were age, sex, race/ethnicity, years of education, employment, and health insurance. Clinical factors assessed were pain-related condition, acute versus chronic condition (≥3 months), perception of illness,¹⁹ physical disability,²⁰ initial symptom improvement, prior use of acupuncture, treatment attitudes,²¹ expectancy of acupuncture,²² out-of-pocket payment, concurrent treatments, treatment barriers, initial treatment satisfaction, recommendation of follow-up treatment, and scheduling of a follow-up appointment at the initial treatment. Age, race/ethnicity, initial treatment effect, and treatment plan correlate with acupuncture use and attendance to follow-up treatment.^6,9,23–25

Pain Symptoms

Pain scores were measured using the VAS (0–100) at 4 time points: day 0 (before and after the initial treatment), day 10, and day 30.²⁶ For pain disability, we used the validated PDI, a 7-item scale scoring from 0 (no disability) to 10 (worst disability) with a possible sum score range of 0–70.²⁷ The PDI was assessed at 3 time points: day 0 (after treatment), day 10, and day 30, with Cronbach α = 0.92 on day 0.

Statistical Analysis

Distributions of all study data were checked for normality and influential outliers, and none was detected. Both intent-to-treat and as-treated analyses (i.e., responder of the study intervention) were conducted using multivariable logistic regression models to compare differences in proportions of acupuncture follow-up treatment returns by study group. The regression models controlled for 2 factors that showed significant group differences at baseline (i.e., having a pain condition and symptom improvement after the initial treatment). Analyses were performed in Stata 13.²⁸ In secondary analyses, all baseline characteristics were modeled to predict follow-up treatment return (yes vs. no) using multivariable logistic regression model and count of follow-up visits using multivariable Poisson regression model. The stepwise forward-selection approach with P-value <0.10 was set in advance to avoid model overfitting relative to sample size.

For repeated measures survey data including VAS and PDI, which was not outcomes of the RCT, the authors analyzed a subsample that reported a pain condition at baseline. First, we conducted a within-subjects test for the slopes of VAS and PDI scores over a 30-day assessment period (days 0, 10, and 30). Then, we tested the effect of those change scores on treatment return (return vs. not return). A repeated measures mixed model “xtmixed” (linear growth model) in Stata 13 was tested with time as a continuous variable. A last-observation carried-forward imputation was used for missing data in days 10 and 30. VAS and PDI models controlled for study group and acute versus chronic pain condition.

Power calculation indicated that it would require >300 participants per group to detect a 10% difference in the primary outcome (i.e., 60% exposure vs. 50% control for treatment return based on preliminary data from Lam⁶). The sample size for this initial pilot study was set to 120 (40 per group) based on the availability of study time and resources.

Results

Participant Characteristics

Participant flow through the pilot trial is shown in Figure 1. After being screened, 120 adults were enrolled for the trial and randomized to 1 of the 3 study groups. In the voice telephone call group, 23 of 40 (58%) patients answered the call. In the text message group, 31 of 40 (78%) patients replied to the text as prompted. No participants asked to not be contacted after receipt of the study exposure.

FIG. 1.

Flow of randomized participants: patient return to follow-up treatment. Primary outcome of return to follow-up treatment was obtained through clinic chart review. All 120 randomized participants had this outcome data available.

Table 1 gives baseline characteristics of the study sample. The mean age of participants was 42.6 years (range 20–80) and 72.5% were female. The mean VAS (before treatment at baseline) was 52.9 ± 25.0. Ninety percent of participants recalled receiving a recommendation from their acupuncturist to return for a follow-up visit. Only 62.5% of the recommendations were written in the site medical record (used in analysis), whereas others were reportedly verbally communicated. Two baseline factors, pain condition (P = 0.03) and initial symptom improvement (P = 0.01), differed by group, and they were included as covariates.

Table 1.

Baseline Characteristics for Participants in the Telephone Call, Text Message, and Control Groups (N = 120)

Characteristics	Telephone call (n = 40), mean ± SD/n (%)	Text messaging (n = 40), mean ± SD/n (%)	Control (n = 40), mean ± SD/n (%)	P
Female	28 (70.0)	28 (70.0)	31 (77.5)	0.69
Age (years)	44.1 ± 13.0	41.0 ± 11.9	42.7 ± 14.2	0.58
White/Caucasian	26 (65.0)	32 (80.0)	29 (72.5)	0.32
≥16 years of school	32 (80.0)	27 (67.5)	26 (65.0)	0.29
Currently employed	35 (87.5)	35 (87.5)	33 (82.5)	0.76
Had health insurance	38 (95.0)	35 (87.5)	36 (90.0)	0.50
Chronic condition (vs. acute)	28 (70.0)	24 (60.0)	23 (57.5)	0.47
Pain condition	20 (50.0)	31 (77.5)	22 (55.0)	0.03
Symptom VAS: before treatment (0–100)	49.9 ± 26.3	53.1 ± 23.5	55.6 ± 25.6	0.60
Symptom VAS: after treatment (0–100)	34.2 ± 23.8	31.4 ± 16.9	39.4 ± 23.9	0.25
Symptom improvement (1–4)a	2.8 ± 1.2	2.1 ± 1.1	2.1 ± 1.1	0.01
Brief illness perception (0–80)	42.8 ± 9.8	44.2 ± 8.4	40.4 ± 10.9	0.22
Physical disability (0–48)	20.4 ± 7.9	21.7 ± 7.1	19.7 ± 8.3	0.50
Attitude toward CAM (0–24)	20.1 ± 4.9	19.4 ± 4.9	20.3 ± 4.6	0.67
Acupuncture expectancy scale (4–20)	16.0 ± 3.5	16.2 ± 2.7	16.6 ± 3.0	0.65
Prior acupuncture experience	30 (75.0)	27 (67.5)	30 (75.0)	0.69
Acupuncture: paid out of pocket	36 (90.0)	38 (95.0)	37 (92.5)	0.70
Concurrent treatment(s)
MD or psychologist visits	15 (37.5)	9 (22.5)	12 (30.0)	0.34
Prescription medications	10 (25.0)	9 (22.5)	7 (17.5)	0.71
Complementary health approaches	26 (65.0)	28 (70.0)	31 (77.5)	0.47
Follow-up treatment barriersb
Time	10 (25.0)	6 (15.0)	7 (17.5)	0.50
Cost	7 (17.5)	7 (17.5)	7 (17.5)	1.00
Distance	10 (25.0)	7 (17.5)	8 (20.0)	0.70
Baseline treatment satisfactionc	19 (47.5)	17 (42.5)	24 (60.0)	0.27
Recommendation for follow-up visitd	25 (62.5)	27 (67.5)	23 (57.5)	0.65
Follow-up appointment scheduled	27 (67.5)	23 (57.5)	17 (42.5)	0.08
Intent to receive follow-up treatment (0–4)	3.6 ± 1.0	3.5 ± 1.0	3.5 ± 0.9	0.92

These items were completed right after treatment on day 0, except the before treatment VAS during study screening. Parentheses show scale sum score range. Control group: treatment as usual with no reminder.

^aSymptom improvement immediately after initial treatment (1: no improvement, 2: some improvement, 3: clear improvement, 4: cure).

^bParticipant reporting 4 or 5 out of a 5-point scale on levels of barrier, 4: quite a bit, 5: very much.

^cParticipant reporting 10 out of a 10-point scale, 10 is completely satisfy.

^dRecommendation for follow-up visit rates based on participant receiving recommendations on paper.

CAM, complementary and alternative medicine; MD, medical doctor; SD, standard deviation; VAS, visual analog scale.

Follow-Up Treatment Return by Group

Of the enrolled sample, 56.7% (68 of 120) of participants returned to a follow-up treatment. Of those who returned, 29.2% (35 of 120) had 1 follow-up visit recorded and 27.5% (33 of 120) had 2–5 follow-up visits recorded in their chart. The adjusted proportion of follow-up treatment returns was 56.3% for the telephone call group, 57.3% for the text message group, and 57.0% for TAU (P = 0.98 and P = 0.99, respectively, compared with TAU) (Fig. 2). The adjusted proportion of follow-up treatment returns in the as-treated analytic sample was 57.0% for the telephone call group and 66.0% for the text message group (P = 1.00 and P = 0.44, respectively, compared with TAU). Only the text message group in the as-treated analytic sample reached a small effect size (π = 0.1) compared with TAU.

FIG. 2.

Adjusted proportion of follow-up treatment return by group assignment. The figure shows the adjusted proportion of patient return to follow-up treatment within 30 days after the initial treatment session at day 0. The number of participants in group is shown at the bottom of each bar. There was no significant group contrast in the intent-to-treat and the as-treated samples. The as-treated analysis focuses on participants who received the intervention. The model controlled for pain condition and immediate symptom improvement.

Predictors of Follow-Up Treatment Return

Table 2 gives the results of 2 regression models. In the logistic regression model, a follow-up appointment scheduled at the initial treatment predicted follow-up treatment return (odds ratio: 5.87, 95% confidence interval [CI]: 2.21–15.59, P < 0.01). In the Poisson regression model, a follow-up appointment scheduled at the initial treatment also predicted a higher count of follow-up visits (incidence rate ratio: 1.95, 95% CI: 1.18–3.20, P < 0.01). Results of the bivariate analyses used for model selection are given in Supplementary Table S1.

Table 2.

Predicting Follow-Up Treatment Return and Number of Follow-Up Visits (N = 120)

Characteristics	Return to follow-up (y/n)		Number of follow-up (count)
	Adjusted OR (95% CI)	P	Adjusted IRR (95% CI)	P
Study intervention
Telephone call	0.90 (0.28–2.87)	0.85	0.84 (0.53–1.33)	0.46
Text messaging	0.50 (0.15–1.75)	0.28	0.84 (0.54–1.31)	0.45
White/Caucasian			1.90 (1.26–2.88)	<0.01
≥16 years of school	2.45 (0.83–7.28)	0.11	1.60 (0.96–2.64)	0.07
Symptom VAS: before treatment (0–100)			1.00 (0.99–1.01)	0.70
Brief illness perception (0–80)			1.03 (1.01–1.14)	<0.01
Acupuncture expectancy scale (4–20)	1.05 (0.90–1.22)	0.55	1.07 (1.00–1.14)	0.04
Acupuncture: paid out of pocket	0.02 (0.00–0.27)	<0.01	0.74 (0.50–1.10)	0.14
Follow-up treatment barriersa
Time			0.83 (0.69–1.01)	0.07
Distance			1.00 (0.82–1.22)	1.00
Recommendation for follow-up visitb	1.85 (0.75–4.60)	0.18	1.14 (0.75–1.77)	0.55
Follow-up appointment scheduled	5.87 (2.21–15.59)	<0.01	1.95 (1.18–3.20)	<0.01
Intent to receive follow-up treatment (0–4)	3.12 (1.39–7.02)	<0.01	1.53 (0.97–2.42)	0.07

We performed adjusted logistic regression in the return to follow-up model, and adjusted Poisson regression for the number of follow-up model. Factors included in each model are displayed with the adjusted OR. Results of the bivariate analyses to inform the model selection are shown in Supplementary Table S1.

^aParticipant reporting 4 or 5 out of a 5-point scale on levels of barrier, 4, quite a bit; 5, very much.

^bRecommendation for follow-up visit based on participant receiving recommendations on paper.

CI, confidence interval; IRR, incidence rate ratio; OR, odds ratio; y/n, yes/no.

Association Between Follow-Up Treatment Return and Pain Trajectory

Of the 73 participants reporting a pain-related complaint at baseline, 43 (58.9%) returned for ≥1 follow-up treatment visits during the 30-day assessment period, whereas 42 (57.5%) responded to at least 1 of the 2 follow-up questionnaires on days 10 and 30. For the pain scale, both returned and not returned participants showed significant improvement in VAS scores immediately after initial treatment at baseline (returned: 62.0 to 34.0, not returned: 54.2 to 34.7, P < 0.01) (Fig. 3a). However, there was no significant improvement across the 30-day assessment period after the baseline treatment improvements nor any between-group differences in change of pain scores during this time (β = −1.8, 95% CI: −5.6 to 1.9, P = 0.33).

FIG. 3.

(a, b) Adjusted VAS and PDI among pain participants for 30 days after initial treatment at baseline, by return to follow-up treatment (n = 73). For (a), participant's pain levels were measured using the visual analog scale (0–100). Group by time difference was nonsignificant (P = 0.33). However, both groups presented a significant improvement in VAS immediately after the initial treatment (P < 0.01). For (b), pain disability was measured by the sum score of the PDI (0–70). Participants who returned to follow-up treatment demonstrated a significant improvement across 30 days (P < 0.01), whereas those not-returned did not (P = 0.54). The result showed a significant group by time difference (P = 0.02). Participant mean score is presented at each time point. Both models controlled for study intervention and acute versus chronic condition at baseline. PDI, pain disability index; VAS, visual analog scale.

For pain disability, participants who returned to follow-up treatment showed a significant reduction in PDI scores at day 30 (28.1 to 20.7, linear β = −3.7, 95% CI: −5.3 to −2.1, P < 0.01) (Fig. 3b). Those who did not return showed a small initial improvement in PDI from days 0 to 10 (22.2 to 19.7, P = 0.19), but gains declined from days 10 to 30 (19.7 to 21.0, P = 0.49). The overall change in PDI scores over the 30-day assessment period was significantly greater in the return group than in the not return group (β = −3.1, 95% CI: −5.6 to −0.6, P = 0.02).

Discussion

This pilot RCT compared the effects of a 1-time voice telephone call reminder, a 1-time text message reminder, and TAU (i.e., no reminder) on patient return to acupuncture follow-up treatment. The findings did not provide evidence that the telephone call or text message reminder increased follow-up treatment return within 30 days. Presence of a follow-up appointment scheduled at initial visit significantly predicted follow-up treatment return. Participants returned for follow-up treatment reported greater improvement in pain disability over a 30-day period. Although the findings did not support the effect of a 1-time telephone call or text message reminder to improve patient return to follow-up treatment, administrative process such as scheduling appointment dates at initial consultation may improve treatment return, and return attendance appears protective of pain disability in the short term.

The study telephone call and text message reminders failed to increase patient return to acupuncture follow-up treatment. The study adopted a minimal 1-time reminder approach, in contrast to the multiple reminder strategy employed by others.^13,14 For example, Arora¹⁸ sent participants text message reminders at 7, 3, and 1 day before their follow-up appointments. However, a 1-time reminder may be more practical for private clinics because it is not labor intensive and more cost efficient. Nevertheless, the intervention message in this study was not specific enough to account for a differential effect between patients with or without an existing appointment to trigger treatment return behavior. We confirmed this alternative explanation by conducting a post hoc analysis looking at the interaction between study intervention group and appointment scheduled on treatment return, and the result was not statistically different. Future research should focus on the use of multiple reminders, combining reminders with treatment plans documented at baseline, and tailor messages to be more refined in their directive of behavioral action (e.g., call to schedule your appointment in the next 7 days).⁶ In secondary analysis, we found that the presence of a follow-up appointment scheduled in the initial treatment predicted follow-up treatment return. Appointment scheduling is a type of contract between patients and the providers.²⁹ Patients in this case are more committed to return for a future visit likely due to their intent to continue care. However, Barlow⁹ suggested that acupuncture patients re-evaluate treatments on an on-going basis. Thus, their intent to continue care may change based on the benefits they receive after each treatment session. For example, patients might stop treatment due to a lack of perceived benefits or gaining the anticipated result.¹⁰ Future research might assess perception in this way and tailor reminders to perception (e.g., benefits sometimes are not immediate and accrue over time, and resolved symptoms may re-emerge over time if not maintained).

A meta-analysis by MacPherson³ found that patient pain symptom was further improved for every 5 acupuncture sessions they received. The current finding supports this by showing that participants who returned for follow-up treatment had greater improvement in pain disability over 30 days. International experts⁴ and historical literature in acupuncture³⁰ cite the use of an extended treatment regimen for conditions such as low back pain (i.e., median 10 sessions). Li³¹ provides biological evidence to show the cumulative effects of prolonged repeated acupuncture stimulations. The null findings for pain symptoms in our study may be due to improved pain scores immediately after the first treatment. Our observed improvements in pain disability support results showing acupuncture may improve pain-related functional abilities over time.^2,32

Study Limitations

This study has several limitations. First, the sample size was not sufficiently powered to detect an effect for the main outcome and so this is a first stage pilot study to uncover necessary effect sizes. Second, recruitment took place at a single acupuncture teaching clinic, which limits the generalizability of the study findings. Third, the study relayed the same message to all voice and text message study groups regardless of whether they held an existing appointment. This could reduce the true effect of a study intervention by not differentiating the subgroups, and there was the missed opportunity to focus our messages on patients either attending an existing appointment or scheduling a new appointment. Fourth, the online questionnaire at day 10 may solicit as an unintended “reminder” to all study participants about their existing treatment, and might trigger return behavior even in the control group. Finally, the study used carried-forward imputation for missing data at days 10 and 30 for the repeated measures, which can bias the treatment effect and underestimate the variance term as imputed values yield no change from the previous time point.

Conclusions

Adding a 1-time telephone call or text message reminder to standard clinic practice did not improve patient return to acupuncture follow-up treatment within 30 days of initial treatment. Scheduling appointment dates at initial treatment may improve treatment return, and return attendance appears protective of pain disability in the short term. Improving acupuncture follow-up treatment return with more tailored messaging is warranted for future research studies.

Author Disclosure Statement

No competing financial interests exist.

Funding Information

No funding was received for this article.

Supplementary Material

Supplementary Table S1