Real World Characterization of Chronic Pain, Success Rates and Implant Rates: Evidence from a Digital Health Platform of Patients Undergoing Spinal Cord Stimulation Evaluations

Abstract

Spinal cord stimulation is an effective treatment for those experiencing chronic back and leg pain but requires a temporary evaluation period (SCSeval) before permanent implantation. We present real-world data from 7000 patients who underwent SCSeval while utilizing a mobile digital health platform for education, feedback, and outcomes collection during their surgical journey. We analyzed preoperative patient demographics, characterized patient pain profiles using the PROMIS-29 surveys, and calculated the rates of conversion from temporary to permanent SCS implantation. Between August 1, 2021 and March 2, 2023, 7000 patients (mean age 59.1, 59.6% female) underwent SCSeval procedures while utilizing a mobile application. Patients commonly experienced aching, sharp, stabbing, tingling, numb, and burning pain. Patients had tried multiple prior therapies and wanted to reduce their use of opioids and pain medications. Overall, 90.1% of patients had a successful SCSeval, and 80.4% of those converted to permanent implant, with the highest rates among those who underwent SCSeval in a hospital setting. There was significant improvement in all domains of pain as evaluated by pre-and-postoperative PROMIS-29 surveys. This study supports the use of digital health technology as part of the SCS journey to improve the patient experience and allow for robust patient reported outcomes collection. The overall rate of SCSeval to permanent SCS in our study of 72.4% was higher than national rates of 64%, suggesting that an app may allow clinicians to better quantify changes in chronic pain and provide more insight into choosing to implant SCS permanently.

Introduction

It is estimated that 20.4% of US adults (50 million) suffer from chronic pain, with 8% of US adults (19.6 million) experiencing severe chronic pain that substantially interferes with their daily lives ¹. Additionally, chronic pain contributes to an estimated $560 billion each year in direct medical costs, lost productivity and disability programs². Spinal Cord Stimulation (SCS) is a proven therapy option for these patients where other medical and surgical therapies are not providing adequate pain relief. Stimulator leads are guided into the epidural space and attached to an implanted pulse generator that provides electrical stimulation to the dorsal horns of the spinal cord, interrupting pain signals³. Clinical trials have proven the efficacy of SCS in chronic lower back and leg pain, failed-back-surgery syndrome, complex regional pain syndrome, and painful diabetic neuropathy^3,4.

Using a temporary SCS device, patients can assess the efficacy of SCS therapy during a 3-10 day evaluation in which clinicians closely observe their patients and monitor their pain levels ³. Typically, the implanting physician defines a successful evaluation as a 50% or higher reduction in pain or an improvement in quality of life. Patients are eligible for a permanent SCS implant if other treatment modalities (such as medication, surgery, physical therapy, or psychological therapy) have failed, they have undergone physical and psychological evaluation, and they have completed a successful SCS temporary evaluation (SCSeval).

Chronic pain experiences are complex and unique to each patient. Pain affects numerous facets of life, including psychosocial well-being, physical activity, and the capacity to work ⁵, and not all these facets can be assessed accurately during a clinic visit. Digital health platforms, or mobile applications, can allow clinicians and patients to track patient-reported outcomes (PROs) over time, give patients the opportunity to self-manage pain, and educate patients about treatment options and surgical recovery, all without the need to visit a medical clinic^6,7. In particular, the SCSeval period could benefit from the ability to monitor real-time pain data and provide feedback, allowing care teams to adjust therapies more efficiently as needed. This is an important step in the pursuit of further enhancing and optimizing the SCS experience and patient-centered neuromodulation.

We present real-world data from a large cohort of patients who underwent SCSeval procedures using a digital health platform for patient education, feedback, and PROs collection throughout the procedural journey. We conducted an observational study in which we analyzed patient demographics, baseline pain characteristics, SCSeval success rate, permanent implant rate, and functional improvement following temporary SCS evaluation.

Methods

Study Design

We performed a retrospective cohort analysis of patients who underwent Spinal Cord Stimulation procedures using the mobile digital health platform CareGuidePro^™ (CGP, developed by Higgs Boson Health for Medtronic, USA).

Description of Digital Health platform

CareGuidePro^™ (CGP) is a digital health platform available on both iOS and Android platforms that provides educational resources, timely notifications, and reminders to patients throughout the temporary evaluation and permanent procedures of receiving SCS therapy. Care providers invite patients to download the app on their smartphones before their SCSeval (Figure 1). Patients can access educational resources, such as Frequently Asked Questions (FAQs) and videos, via the app. They can also communicate with their device team via an integrated messaging function. The CGP app also allows patients to complete standard and individualized PRO surveys. In addition, the platform provides secure/encrypted sharing of these PRO data with a patient’s physician and care-team to keep them informed of the patient’s status.

Figure 1: Timeline for Engaging with Digital Health Platform.

Timeline for engaging with the digital health platform from App download to End of temporary SCS Evaluation Period (SCSeval). Baseline surveys are sent before the SCSeval to gather information about history of Chronic pain, and the PROMIS-29 survey characterizes 7 domains (pain interference, ability to participate in social roles and activities, sleep disturbance, fatigue, depression, anxiety and physical function)

The patient receives a baseline pain history survey (Supplementary Material 1) and PROMIS-29 (Patient-Reported Outcomes Measurement Information System^®) survey after enrolling in CGP and prior to the SCSeval^8–10. (Supplementary Material 2). Patients are also given the PROMIS-29 survey at the conclusion of the evaluation period.

The PROMIS-29 profile measure assesses seven domains (pain interference, ability to participate in social roles and activities, sleep disturbance, fatigue, depression, anxiety and physical function) and pain intensity. Higher PROMIS symptom scores reflect worse symptom burden, and higher PROMIS function scores reflect better functioning. The PROMIS domain scales are scored on a T—score metric with a mean of 50 to represent the average in the US general population and a standard deviation of 10. The exception is sleep disturbance with a score of 50 representing the average in a mix of those from the US general population and clinical sample experiencing sleep problems.

At the conclusion of the SCSeval, a feedback survey is collected to characterize the user experience and utility of the platform (Supplementary Material 3). This manuscript is based on Pain History, PROMIS-29, and feedback survey data collected during the SCSeval period that lasts between 3 and 10 days.

Study Cohort & Data Source

Patients who received a SCSeval between August 1, 2021 and March 2, 2023 were included in the study cohort. These patients were treated by 2,654 implanting physicians at 1,007 institutions in the United States. These patients were invited to download CGP and subsequently signed up for the app. A random sample of 7000 patients from this cohort was used to characterize their pain history and SCS therapy response. The randomization of this patient sample was performed before any statistical analysis was begun. As depicted in the cohort diagram, specific endpoint analyses were conducted on subgroups of this cohort (Figure 2). Patients who completed the Baseline Pain survey were categorized according to their chronic pain. The functional enhancement of patients who responded to the PROMIS-29 survey at baseline and at the conclusion of SCSeval was evaluated. Patients with the requisite datapoints had their SCSeval success rate and permanent implant rate evaluated.

Figure 2: Cohort Diagram of CGP Patients undergoing SCS Evaluation.

Cohort diagram illustrating inclusion and exclusion of patients, as well as those who completed SCS evaluations and PROMIS-29 surveys.

° ASC = Ambulatory Surgical Center. Defined as a freestanding facility, other than a physician’s office, where surgical and diagnostic services are provided on an ambulatory basis.

°° Office. Defined as location, other than a hospital, skilled nursing facility (SNF), military treatment facility, community health center, state or local public health clinic, or intermediate care facility (ICF), where the health professional routinely provides health examinations, diagnosis, and treatment of illness or injury on an ambulatory basis

°°° Hospital. Defined as a facility, other than psychiatric, which primarily provides diagnostic, therapeutic (both surgical and nonsurgical), and rehabilitation services by, or under, the supervision of physicians to patients admitted for a variety of medical conditions.

* Unknown

Data Source

In this retrospective cohort observational study, two different data sources were utilized. Manufacturer records (Medtronic plc, Minneapolis, Minnesota, United States) included SCSeval procedure date, outcome designation (success/failure), permanent implant date, patient age, gender, implant indication, SCSeval stimulation type, and site of service. The CGP app data contained survey responses from the duration of the SCSeval. To facilitate cohort analytics, these two sources were linked with a common identifier present in both datasets. The study was a retrospective analysis of deidentified data, collected by a HIPAA-exempt provider of health care in the U.S. Patient consent for data use was obtained as a part of the disclosure when first downloading the app. These data were not collected as a part of a clinical study, and thus are exempt from institutional review board approval under applicable law. The data was processed in accordance with applicable law and deidentified so that only columns noting age, gender, implant indication, trial type, site of service, trial outcome, conversion rate from both successful and unsuccessful trials, and trial-to-implant time were present.

Data Processing/Variable Transformation

Real-world data consists of numerous absent datapoints and the possibility of variable data entry. The cohort diagram (Figure 2) illustrates the missing datapoints and the cohorts used for specified analyses. The data were processed further to eliminate duplicates. The study cohort comprised of de-novo SCSeval, meaning patients with prior SCSeval were excluded from the study cohort. In order to account for outliers in entry of continuous variable (age), we applied a 1% trim to this data field. This meant, we removed the top 0.5 percentile and the bottom 0.5 percentile of the distribution to derive the age statistics.

Statistical Analysis

All analyses were performed using R version 4.2.1, (R Foundation for Statistical Computing, Vienna, Austria) and Python 3.7 programming language (Python Software Foundation. Python Language Reference, version 3.7. Available at http://www.python.org). Alpha was set at 0.05 with adjustments for multiple comparisons.

Chronic Pain Classification

To quantify chronic pain, a tabulation of patient characteristics at baseline (prior to SCS therapy) was utilized. The pain history questionnaire facilitates this categorization by detailing the pain’s location, description, duration, treatments attempted, impact on daily life, and opioid use. Summary statistics were also compiled for patient characteristics (age, gender), SCSeval indication, SCSeval type attempted, and location of service.

We wanted to reveal sets of pain characteristics that are more likely to co-express aiming to trace predominant causes of chronic pain and how it evolves in time. We applied hierarchical clustering¹¹ to the pain characteristics, where each characteristic was represented by the set of patients reported on it. The distance between two characteristics was defined as Jaccard dissimilarity score¹² between those sets. Intuitively, two characteristics are close if they appear together more frequently in the same survey than they do individually.

The distance between two clusters was defined as the maximum distance between all pairs of characteristics, such that each characteristic belongs to a distinct cluster. Clusters were identified by pruning the resulted hierarchical tree using threshold of 0.5 This guarantees that any two characteristics within a cluster intersect in at least fifty percent of patients. Separate clusters of pain characteristics were calculated for patients who reported experiencing pain for 6-12 months, 1-3 years, and over 3 years.

SCSeval Success Rate

At the conclusion of each SCSeval, the care provider determined whether the SCSeval was successful or unsuccessful based on pain reduction and observed changes in the patient. The success rate for the entire cohort was determined, and then stratified by site of service. The sites of service were, as defined by the US Centers for Medicare & Medicaid Services, ambulatory service center (non-office, outpatient center where procedures are performed), in-hospital (inpatient facility) and in-office (physician’s office/clinic).¹³ The comparison across the sites of service were performed using pairwise proportion comparison test. The p-values were adjusted using false discovery rate (FDR) ¹⁴.

Implant Rate

The rate of permanent SCS implant after a successful SCSeval was observed over an 8-month time period with a minimum 2-month follow-up period after SCSeval. This meant that the implant rate was observed in a subset of patients who received their SCSeval between August 1, 2021 and December 31, 2022, with follow-up through March 2, 2023. (Figure 2). Over the course of the 8-month follow-up, cumulative implants were determined as the proportion of patients implanted over time; the rate was derived at 3 and 6 months after the trial. In addition, the hazard rate (p/day) was calculated using a bin length of 7 days to evaluate the change in the probability of receiving a permanent SCS implant based on the absence of an implant up until that date. This analysis was additionally stratified by service location (ambulatory service center (ASC), in-hospital, and office).

Functional Improvement

Using the PROMIS-29 instrument, the functional improvement at the conclusion of the SCSeval relative to the pre-trial baseline was measured. PROMIS-29 has been validated for measuring chronic pain in the surgical and back pain populations.^9,15,16. Assessment Center uses a T-score metric to score PROMIS-29 scales^17,18 (see www.assessmentcenter.net and https://www.youtube.com/watch?v=KM2FqYo--A). Each domain’s PROMIS T-Score is a standard score with a mean of 50 and a standard deviation of 10. Each domain’s T-score metrics were compared between time points using a paired sample t-test. The T-scores for each domain are categorized as normal, mild, moderate, or severe according to predefined ranges.^19–21 At pre-trial baseline and trial conclusion, the proportion of patients in each category for each domain was evaluated.

Quantifying the Utility of the Digital Health Platform

At the conclusion of the SCSeval, a feedback survey was administered to quantify the utility of this platform in preparation for the procedure, recovery, and app usability. We also assessed descriptive statistics for the time required to complete the baseline pain questionnaire and the PROMIS-29 instrument.

Results

Patient Population and SCSeval Details

Between August 1, 2021 and March 2, 2023, 7000 patients utilized the CGP mobile application to undergo SCSeval procedures (Figure 1). These patients comprised a nationally representative sample, receiving treatment from 2654 implanting physicians at 1007 U.S. institutions. At the time of SCSeval, the average patient age was 59.1±13.7yrs, with 39.7% of patients being above the age of 65. Of 5912 patients with known sex, 3522 (59.6%) were female and 2390 (40.4%) were male. Geographically, 1449 (20.7%) of enrolled patients were from the Central U.S., 1547 (22.1%) from the Great Lakes, 1022 (14.6%) from the Northeast, 2058 (29.4%) from the Southeast, and 931 (13.2%) from the West.

Chronic low back pain (n=2310, 33.0%) was the most prevalent indication for SCSeval, followed by radiculopathy (n=654, 9.3%) and degenerative disc disease (n=210, 3.0%). The type of stimulation parameters used in these trials was Differential Target Multiplexed (DTM^™ – a program of multiple signals with varying pulse widths, amplitudes, and frequencies) in 67.1% of patients, DTM^™ Endurance (low-energy version of DTM) in 9.3% and Evolve (Low and High-Density 1kHz pulses) in 6.5%^22,23. The characteristics of these patients are detailed in Table 1. The service location for the SCSeval was Ambulatory Service Center for 2117 (30.2%) patients, office for 1877 (26.8%) patients, and Hospital for 2084 (29.8%) patients.

Table 1.

Characteristics of Patients Enrolled in a Digital Health Platform for SCSeval.

Characteristic (N=7000)	Statistic
Age (years)	59.1 ± 13.7 [IQR 50 70]
% Pts with Age ≥ 65yrs	39.7%
Sex Female N (%)	3522 (50.3%)
Sex Male N (%)	2390 (34.1%)
Sex Unknown N (%)	1088 (15.5%)
Geographic Region, n (%) ^
Central	1449 (20.7%)
Great Lakes	1547 (22.1%)
Northeast	1022 (14.6%)
Southeast	2058 (29.4%)
West	924 (13.2%)
SCSeval indication, n (%)
Back Pain*	2310 (33.0%)
Chronic Regional Pain Syndrome (CRPS)	147 (2.1%)
Herniated Disc**	210 (3.0%)
Radiculopathy	654 (9.3%)
Unknown/Other***	3475 (49.6%)
Diabetic Painful Neuropathy	204 (2.9%)
SCSeval Type, n (%)
DTM	4695 (67.1%)
DTM Endurance	650 (9.3%)
Evolve	459 (6.5%)
Unknown	1197 (17.1%)
SCSeval Site of Service n (%)
Ambulatory Service Center (ASC)	2117 (30.2%)
Office	1877 (26.8%)
Hospital	2084 (29.8%)
Unknown*	922 (13.2%)

^*Back Pain = Chronic Low Back Pain, Failed Back Surgery Syndrome, FBSS Leg/Back

^**Herniated Disc = DDD/Herniated Disc, Degen Disc Disease / Herniated Disc Pain

^***Unknown = (Left empty), Failed Back Surgery Syndrome, Upper Limb Pain, Other

^{^}Regions defined as follows. Central: North Dakota, South Dakota, Nebraska, Kansas, Oklahoma, Texas, Iowa, Missouri, Arkansas, Louisiana, Tennessee, West Virginia. Great Lakes: Illinois, Indiana, Michigan, Minnesota, Ohio, Wisconsin. Northeast: Connecticut, Maine, Massachusetts, New Hampshire, New Jersey, New York, Pennsylvania, Rhode Island, Vermont. Southeast: Alabama, Florida, Georgia, Kentucky, Mississippi, North Carolina, South Carolina, Tennessee, Maryland, District of Columbia, Virginia, Delaware, and Washington D.C. West: Arizona , Colorado, Idaho, New Mexico, Montana, Utah , Nevada, Wyoming, Alaska, California, Hawaii, Oregon, Washington.

Overall, 2978 patients completed the pre-procedural Baseline Pain History survey, 3291 completed the Baseline PROMIS-29 survey, and 1861 completed both baseline and end of SCSeval PROMIS-29 surveys. Overall, patients who completed surveys were younger (Pain History: 57.3 vs 60.3 years old; PROMIS-29: 57.4 vs 60.1 years old). More details regarding survey completion demographics can be found in Supplementary Material 4.

Pain History

Prior to undergoing the SCSeval procedure, 2,978 (42.5%) patients completed the Baseline Pain History survey, which characterized pain location, description, duration, prior treatments attempted, impact on life, and opioid use. (Table 2). Location and description of persistent pain were observed to be multidimensional. 66.1% of respondents reported pain lasting > 3 years, and 25.3 % reported pain lasting between 1 and 3 years. The most common pain descriptions were aching (75.6%), sharp (69.2%), tingling/numbness/pins & needles (59.7%), shooting (55.4%), stabbing (54.9%), and scorching (54.7%). Pain most commonly affected the lower back and legs. Prescription medications (88.21%), heat/ice packs (87.8%), targeted injections (84.5%), and therapy, acupuncture, or massage (75.8%) were among the prior treatments performed. The majority of patients who used analgesics wished to reduce their dosage. The majority of patients reported that pain affected their ability to conduct daily tasks, exercise, sleep, work, and interact with others.

Table 2:

Baseline Pain History

Characteristic (N=2978)	n	%
Pain Location
Upper Limb (arms/hands)	379	12.73 %
Other location	737	24.75 %
Lower Limb (feet/ankles)	865	29.05 %
Right Leg	1602	53.79 %
Left Leg	1646	55.27 %
Lower Back	2561	86.0 %
Pain Description
Other pain	220	7.39 %
Prickly	593	19.91 %
Difficult to describe	632	21.22 %
Burning	1626	54.60 %
Stabbing	1634	54.87 %
Shooting	1650	55.41 %
Tingling/numbness/pins and needles	1779	59.74 %
Sharp	2061	69.21 %
Aching	2251	75.59 %
Pain Duration
Less than 6 months	31	1.04 %
6-12 months	224	7.52 %
1-3 years	754	25.32 %
3 years or longer	1969	66.12 %
Treatments Tried
Chiropractic Care	516	17.33 %
Other treatment	578	19.41 %
Topical pain relievers	2123	71.29 %
Over the counter medication	2208	74.14 %
Therapy, acupuncture, massage	2257	75.79 %
Targeted injections	2515	84.45 %
Heat/Ice Pads	2616	87.84 %
Prescription medications (opioids, etc.)	2627	88.21 %
Impact of Pain on life
Ability to socialize	2101	70.55 %
Ability to work	2176	73.07 %
Mood	2283	76.66 %
Ability to sleep	2439	81.9 %
Ability to perform household chores	2712	91.07 %
Ability to exercise	2718	91.27 %
Opioid Usage	2173	72.97%
Goal of Reducing Opioid usage	2407	80.76 %
Other impact	373	12.53 %

Figure 3 depicts the hierarchical clustering of pain characteristics. Data from patients who reported experiencing discomfort for 6-12 months revealed three clusters: 1) Sharp pain, lower back pain, aching pain, and the use of targeted injections & heat/ice pads; 2) Ability to exercise, ability to perform chores, ability to sleep, ability to work, ability to socialize, mood, prescription medications, over the counter medications, & topical pain relievers; 3) Opioid use and a target to reduce opioid use. At 1 to 3 years of pain, two new clusters emerged: 1) tingling/numbness/pins-and-needles pain and searing pain; 2) shooting pain and sharp pain. At 3+ years of pain, three clusters emerged, with one cluster combining most of the previous clusters and another cluster combining left and right leg pain with tingling/numbness/pins-and-needles pain.

Figure 3: Clusters of Chronic Pain Characteristics using Hierarchical Clustering.

Clusters of pain characteristics at 6-12 months, 1-3 years, and 3+ years among patients who completed pain surveys. These plots were made using Jaccard distance and a complete-link linkage method with a cutting threshold of 0.5.

SCSeval Success

For 5037 patients with adequate follow-up data and a known outcome for their SCSeval, the success rate, indicating adequate pain control to be eligible for permanent implant, was evaluated. The overall success rate was 90.1% (n=4539). Overall success rates were 89.3% for ASC (n = 1605 of 1798 patients), 91.5% for hospital (n = 1554 of 1699 patients), and 89.6% for office (n = 1380 of 1540 patients) when stratified by site of service.

In the six months following a successful SCSeval, a total of 3649 (80.4% of successful SCSeval and 72.4% of 5037 patients) patients underwent conversion to permanent implant. 81.4% (n=1306 of 1605) of ASC patients, 83.7% (n = 1301 of 1554) of hospital patients, and 75.2% (n = 1038 of 1380) of office patients underwent permanent implant placement. (Figure 4). For all three service sites, the SCSeval conversion hazard rate crested within 60 days and then declined.

Figure 4. Permanent SCS Implant Rate.

Plots showing the cumulative SCSeval conversion rate over time, as well as the hazard rate per day after successful SCSeval procedures. The hazard rate (p/day) was derived to evaluate change in probability of receiving a permanent SCS implant conditional on absence of an implant until that date.

PROMIS-29 Profile Survey Results

A total of 1861 (26.6%) patients completed paired PROMIS-29 surveys at the beginning and conclusion of SCSeval. At baseline, the mean T-score of the sample was within the normal range for the domain of depression, within the mild severity range for sleep disturbance, fatigue, and anxiety, and within the moderate range for pain interference, ability to participate in social roles & activities, and physical function. At the conclusion of the trial, the physical function and pain interference domains’ mean T-scores were within the mild severity range, while all other domains’ T-scores were within the normal range. The T-scores between Baseline and Last day of SCSeval were compared via paired t-test and were observed to be statistically significantly different for each one of the seven domains (p < 0.001). In Table 3, the ranges of T-scores are reported and color-coded according to the classifications of normal, mild, moderate, and severe. At each timepoint, stacked bar plots depict the proportion of patients with T-scores in the normal, mild, moderate, and severe ranges. (Figure 5). The pain intensity levels in this cohort of 1861 patients were 7.13 ± 1.67 (median 7 IQR [6-8]) at baseline, and at end of SCSeval were 3.78 ± 2.34 (median 3 IQR [2-5]), p < 0.001. The proportion of patients in the severe category decreased across all PROMIS domains by the end of the SCSeval period, while the proportion of patients in the normal and mild categories increased. (Figure 5)

Table 3.

PROMIS-29 Domain T-Scores Among Patients with Paired Baseline & End of SCSeval Data

PROMIS-29 domains	Baseline			Last Day of SCSeval
	Mean	Mean − SD	Mean + SD	Mean	Mean − SD	Mean + SD
Pain Interference*	68.0	62.4	73.7	58.5	50.6	66.5
Ability to Participate in Social Roles and Activities°	38.7	31.7	45.8	46.0	38.1	54.0
Sleep Disturbance*	58.4	49.9	66.8	48.7	40.2	57.1
Fatigue*	59.2	50.3	68.1	50.5	42.0	59.0
Depression*	54.4	45.0	63.9	49.1	40.8	57.4
Anxiety*	57.2	47.7	66.7	50.7	42.0	59.4
Physical Function°	34.0	29.3	38.8	40.5	33.9	47.2

^°Defined as: Normal (≥45), Mild (40-45), Moderate (30-40) and Severe (<30)

^*Defined as: Normal (<55), Mild (55-60), Moderate (60-70) and Severe (≥ 70)

Severe

Moderate

Mild

Normal

Figure 5. Longitudinal Changes in PROMIS-29 Severity.

Stacked barplot showing the percentage of patients in Severe, Moderate, Mild and Normal PROMIS-29 severity at Baseline and End of SCSeval.

App Response Survey

The feedback on the CGP app was provided by 2197 patients (31.4%). The results showed that 78.0% of users found CGP helpful in preparing for their SCSeval procedure, 79.3% found it helpful during recovery from the procedure, 83.1% found it helpful in answering questions about neurostimulation, 90.6% found it easy to fill the app-based surveys, 91.35% found it user friendly, and 86.4% found the messages & notifications to be helpful throughout the surgical journey. The survey responses were not time-consuming or burdensome. The time to fill the pain baseline survey via the app was a median of 2.2 minutes [IQR 1.6-3.1], and the time to fill the PROMIS-29 instrument was a median of 3.7 minutes [IQR 2.7-5.5].

Discussion

In this analysis of real-world data, we demonstrated the usefulness of a mobile digital health platform (CGP) as a tool for outcome collection during the SCSeval period. In addition, we presented analyses monitoring the chronic pain profiles of SCS recipients and the transition from SCSeval to permanent implant.

In recent years, digital health tools have gained popularity due to their capacity to remotely capture patient data and provide patients with instructions and education outside of the traditional clinic setting. Studies into these platforms have demonstrated high rates of patient satisfaction, with patients reporting high satisfaction in apps for preparing and recovering from spine surgery⁶ and heart surgery²⁴. Digital health platforms have been shown to reduce emergency room visits and hospital admissions in acute postsurgical contexts, as well as assist those with chronic diseases in self-managing their conditions by increasing patient engagement and awareness ^25–27.

While the clinical utility of a screening evaluation prior to permanent SCS is debatable^28,29, a trial procedure demonstrating pain relief is required for insurance approval in the United States³⁰. Studies indicate that pain recall is influenced by the timing of the survey, the current degree of pain, and anxiety levels.^31,32. Single item scales such as the Numerical Rating Scale (NRS) and Visual Analog Scale (VAS) are easy to use in the clinical setting but lack the granularity of multi-item scales such as PROMIS-29, which measure multiple domains impacted by pain including physical, psychological, and social aspects of well-being ^33,34. Thus, digital health tools enable patients to rapidly complete these in-depth surveys and provide pain assessments outside of the clinic, while also allowing providers to view this data during a SCS evaluation and evaluate the device’s efficacy.

The ability to collect real-world data in a clinically seamless fashion is a key benefit of using a digital health platform such as CGP. Although randomized controlled trials (RCTs) are the gold standard for evidence-based medicine, the strict inclusion and exclusion criteria of these studies frequently limit their population generalizability ³⁵. Chronic back pain and neuropathic pain are known to disproportionately impact women and those in their mid-to-late 50s and older, making the patients in this study representative of the general population^7,36,37.

As part of the device and drug approval process, regulatory agencies such as the US FDA, Medicare and commercial payers are increasingly relying on real-world data to validate the results of RCTs ^38,39. We were able to obtain pain history surveys from 2978 (42.5%) patients and longitudinal paired PROMIS-29 data from 1861 (26.6%) patients, representing a robust real-world dataset.

Evaluation of Chronic Pain before & After SCS

This digital-health platform’s data allowed for a precise quantification of chronic pain profiles, highlighting the multidimensional and pervasive nature of pain. Across all three evaluated durations of pain (3-12 months, 1-3 years, and 2+ years), we discovered that pain frequently interfered with the ability to perform activities of daily living, such as exercising, housework, and sleeping, which are well-known to be detrimental effects of chronic back pain ⁴⁰. Patients also reported using recommended guideline therapies for chronic lower back pain, such as physical therapy, opioids, targeted injections, and heat or cold pads ⁴⁰. Concerningly, patients at every timepoint commonly endorsed a desire to reduce opioid use, confirming the nationwide trend of high opioid utilization for lower back pain⁴¹. Mood effects were also common in patients with chronic pain,^42,43 suggesting that earlier intervention with SCS may be warranted to prevent patients from developing opioid use disorders or mood disturbances.

When comparing baseline and post-SCSeval data for the population, PROMIS-29 survey results revealed an improvement in every measured domain. The survey captures the patient experience and functions as a mechanism for care providers to receive feedback. In lieu of relying on single-item pain measures, surveys such as PROMIS-29 give providers a quantitative method for determining if a patient has improved significantly during the trial period. While RCTs have demonstrated the efficacy of SCS in reducing pain in patients suffering from chronic back & leg pain,^44,45, diabetic neuropathy⁴⁶, and failed back surgery syndrome⁴⁷, few studies have quantified the effects of SCS on other aspects of life. Even though our study only measures short-term data from the SCSeval period, our results indicate that the SCS can be efficacious within 3 to 10 days. As demonstrated here, SCS affects multiple aspects of personal and social functioning, and our data support the use of multi-domain patient reported outcomes surveys to examine the full impact of SCS on a patient’s life. Planned future studies will monitor longitudinal PROMIS-29 survey results over longer follow-up periods following permanent SCS implantation.

Permanent Implant Rate

The nationwide rate of conversion from SCSeval to a permanent implant is about 64%^48,49. Higher conversion rates were seen among those who received SCSeval from high-volume providers and from neurosurgeons or orthopedic surgeons compared to anesthesiologists or physiatrists^48,49. In our study, the overall temporary to permanent implant conversion rate in the SCSeval cohort was 72.4%. This is higher than the national rate, raising the question of whether CGP contributed to this outcome. However, we did not include a direct comparison group which didn’t utilize CGP, so this conclusion cannot be made outside the context of a prospective clinical trial with an adequate control group.

The SCSeval success rate was 90.1%, and 80.4% of successful evaluations received a permanent implant, with a very similar rate of successful pain reduction among patients at ASC, Office, and Hospital location. However, our analysis revealed that those who underwent SCSeval procedures in an office-based practice had reduced permanent conversion rates at 3 and 6 months compared to those who underwent a trial at an ASC or hospital.

SCSeval procedures are performed by multiple specialties (including anesthesia and physiatry), but the permanent implant is often done by a different individual (interventionalist or surgeon), who places the final stimulator leads and implantable pulse generator in an operating room ⁵⁰. This suggests that office-based trial procedures may have a barrier that prevents the scheduling of a permanent trial, whether due to insurance considerations or scheduling an appointment with a surgeon to conduct the permanent implant in an operating room at a separate facility. While studies are necessary to confirm this, we can hypothesize that trials conducted in an ASC or inpatient hospital setting were more likely to convert to permanent implant due to the lack of this logistical barrier, as both the SCSeval and permanent implant can be done in operating rooms at the same facility if desired. As the probability of obtaining a permanent implant decreases over time, it is important that patients have access to care and the proper environment to facilitate the implant procedure.

Patient Perspectives on Digital Health

Over 2100 patients provided feedback on the use of CGP, with responses indicating a positive experience with the platform. Patients reported that the platform was user-friendly and useful in guiding them through the SCSeval procedure’s preparation and recovery. The completion of patient reported outcome surveys was not time-consuming or burdensome, as evidenced by short time to fill these surveys. In the context of neuromodulation, digital health tools can provide the ability to effectively engage patients in their therapy journey and enable providers to evaluate the effectiveness of treatment based on patient-reported outcomes.

Limitations & Future Directions

Despite presenting a large sample of patients undergoing SCSeval procedures while using a digital health platform, our study has several important limitations. The first limitation is that this study was limited to a single procedure (SCSeval) and does not contain long-term data beyond the SCSeval period. In addition, this research is limited to the United States and cannot be extrapolated to other nations where SCS indications, health insurance and payer policies may vary. In addition, we lacked data on the sociodemographic characteristics of our population to evaluate the generalizability according to race, ethnicity, and socioeconomic status. We also lacked details on the dosages and types of medications patients used, and thus were not able to more specifically characterize the medication utilization among SCSeval patients. This could be remedied by linking electronic medical record data or administrative claims data in the future.

Another major limitation is that our study did not include a control comparator group that was not enrolled in CGP and thus did not have access to the educational content and surveys found within the app. A future clinical trial could be planned to compare these two cohorts and determine if use of this is associated with differences in SCSeval success, permanent conversion, and patient reported outcomes.

Patients may have been incentivized to complete surveys and interact with the app during the trial period, but we did not evaluate survey responses at multiple timepoints following permanent SCS implantation. Similar to other studies that collected data from digital health platforms, our data were restricted by response bias ⁵¹. Less than half of app users completed baseline pain surveys (47.0%), longitudinal PROMIS-29 surveys (26.6%), and app feedback surveys (31.4%), which may have skewed our results in favor of patients who were more content with the app or who experienced greater pain relief. Additionally, we found that survey responders tended to be younger, with a higher percentage of female responders. It has already been demonstrated that older patients are less likely to engage with mobile and digital health technologies.⁵² Other studies on digital health apps from our study authors have demonstrated PROMIS-29 survey completion rates as low as 30-58% postoperatively,^6,53 suggesting that long-term engagement with digital health tools should be a target for future development. Potential issues leading to lower engagement include a lack of time or incentive to complete surveys, information overload, difficulty using the app.

Personalized reminders for survey and task completion, presenting survey results back to patients, and allowing for direct contact with providers through the application are potential strategies for increasing patient engagement ⁵¹. Additionally, “digital health literacy”, in the form of increasing readability, making more intuitive patient interfaces, offering technical support, and asking for patient feedback, has been proposed in the face of increasing proliferation of digital health tools for healthcare.⁵² A potential solution to address this in our study would be to reach out to patients who didn’t engage with the app and determine what roadblocks exist to engagement with the technology.

Several potential future studies could be derived from this work. To quantify the impact of permanent SCS, we plan to conduct future studies analyzing longitudinal data from patients’ post-implant SCS PROMIS-29 surveys. We additionally anticipate distinguishing the patient experience and pain characteristics of SCS indications, such as failed back surgery syndrome, chronic regional pain syndrome, and painful diabetic neuropathy.

Conclusion

This study presents data from a large cohort of patients who underwent SCSeval procedures using a mobile digital health platform. The results support the use of digital health technology as part of the SCS journey to enhance the patient experience and allow for the robust collection of patient-reported outcomes. We were able to characterize the pain profiles of patients undergoing SCSeval and examine the trial-to-permanent conversion rates. Our findings suggest that using CGP may enable clinicians and representatives to better track and quantify changes in chronic pain, thereby providing more insight into the decision to permanently implant SCS.

In addition, we were able to demonstrate the use of PROMIS-29 as an effective method to characterize the multiple health-related quality of life domains affected by pain, as well as improvement in all domains among SCSeval-treated patients. Therefore, we recommend that digital health tools like the one described here be utilized in all SCS evaluations to enhance patient experiences and data collection during the process of undergoing an SCSeval and deciding to proceed with a permanent implant.

Future research should seek to characterize the long-term pain outcomes of patients with permanent SCS implants and determine how pain etiology influences postoperative outcomes. This characterization of real-world data has the potential to evolve the field of neuromodulation with a better understanding of best practices and the creation of therapy options tailored to the etiology of pain and specific pain subtypes.

Perspective.

This article presents real world evidence from a digital health platform for therapy education and outcomes collection from patients undergoing spinal cord stimulation evaluation procedures. Such tools could allow for better pain characterization and allow for more nuanced tracking of patient outcomes among those with chronic pain.