Home-Based Remote Programming versus Inpatient Programming After Short-Term Spinal Cord Stimulation for Herpes Zoster Neuralgia: A Single-Center Retrospective Cohort Study

Abstract

Objective

Short-term spinal cord stimulation (st-SCS) is increasingly used for acute and subacute herpes zoster neuralgia (HZNP), but postoperative management typically requires prolonged hospitalization. We compared two management strategies—continuous inpatient programming versus early discharge with remote programming—on pain outcomes, patient-reported outcomes, safety, and costs.

Methods

This single-center retrospective cohort study included 28 patients with HZNP (disease duration ≤3 months) who underwent percutaneous st-SCS between May 2021 and May 2024. Fourteen patients received inpatient programming throughout the stimulation period; 14 were discharged 1–2 days post-implantation with remote parameter adjustment. Pain (VAS), sleep (PSQI), anxiety/depression (HADS), quality of life (SF-36), opioid consumption (MME), complications, and costs were assessed preoperatively, at electrode removal, and at 3, 6, and 12 months.

Results

Both groups achieved substantial pain relief after st-SCS (time effect P<0.001). Between-group differences in VAS scores were less than 0.5 points at all postoperative time points—well below the minimal clinically important difference of approximately 2 points—indicating comparable analgesic efficacy regardless of programming setting (group effect P=0.325; group×time interaction P=0.902). At 3 months, the home group had numerically lower PSQI (5.8±1.3 vs 7.1±1.5) and HADS scores (6.2±1.8 vs 7.8±2.1) and higher SF-36 scores (78.4±6.2 vs 72.5±6.8), but these differences did not survive Bonferroni correction. Opioid consumption was lower in the home group at 3 months (13.2±4.0 vs 18.0±4.2 mg/day, P=0.005). Complication rates were similar (14.3% each). Total healthcare costs were 28.5% lower in the home group (9464±938 vs 13,236±1499 CNY, P<0.001), driven primarily by reduced bed and nursing fees from shorter hospitalization.

Conclusion

For patients with HZNP treated with st-SCS, early discharge with remote programming achieved analgesic results comparable to inpatient management—both in statistical terms and, more importantly, in clinical magnitude—while reducing opioid use and healthcare costs. The cost advantage reflects shorter hospital stay rather than superior analgesia. These findings support remote programming as a practical alternative for appropriately selected patients, though confirmation in larger prospective studies is needed.

Introduction

Herpes zoster neuralgia can be a debilitating condition. When pain persists beyond 3 months after rash onset, it is classified as postherpetic neuralgia (PHN), a chronic neuropathic pain state that is notoriously difficult to treat and that substantially erodes sleep, mood, and daily function. The risk of developing PHN rises sharply with age; among patients over 60, roughly half continue to experience pain¹ months after the acute eruption. Patients in the acute or subacute phase (pain duration ≤3 months) represent a window of opportunity—intervening effectively during this period may prevent the transition to chronic PHN, a far harder problem to solve.²

Short-term spinal cord stimulation has gained traction as a minimally invasive option for acute and subacute HZNP. It is important to distinguish this application from the conventional use of short-term SCS as a diagnostic trial to screen candidates for permanent implantation in chronic pain syndromes such as failed back surgery syndrome. In HZNP, st-SCS is itself the treatment: a percutaneous temporary electrode is placed in the epidural space for a limited period (typically 10–17 days) to modulate pain signaling during the period of active neural sensitization. The rationale extends beyond the classic gate-control mechanism;³ contemporary work points to modulation of dorsal horn glial activation, suppression of pro-inflammatory cytokine release, engagement of descending inhibitory pathways, and recruitment of endogenous opioid and cannabinoid systems.

The standard postoperative approach has been to keep the patient in hospital for the entire stimulation period, with daily bedside programming by a physician. This ties up beds, drives up costs, and is inconvenient for patients. Remote programming has been adopted successfully in the follow-up of permanently implanted SCS systems, where it has been shown to maintain efficacy while cutting clinic visits.^4–6 However, temporary st-SCS with externalized leads carries a different risk profile from fully implanted systems—exit-site infection, lead migration, and accidental dislodgement are real concerns during the 10–17 day externalized window. Reported complication rates for temporary SCS range from 5% to 20%. Whether remote programming can be applied safely and effectively in this setting has not been adequately studied.

We conducted a retrospective cohort study comparing two postoperative management strategies for st-SCS in HZNP: continuous inpatient stay with daily bedside programming until electrode removal, versus early discharge with remote parameter adjustment supplemented by outpatient dressing changes and scheduled visits. Both groups underwent identical in-hospital implantation and initial programming; the strategies differed in where and how subsequent adjustments were made, and consequently in length of stay, nursing intensity, and patient environment. We examined analgesic efficacy, patient-reported outcomes, safety, and costs, interpreting the results through a clinical lens rather than relying solely on statistical thresholds.

Materials and Methods

Study Design and Participants

This single-center retrospective cohort study was conducted at the Department of Pain Medicine, The First People’s Hospital of Yibin, Sichuan, China. We reviewed the medical records of all consecutive patients who underwent st-SCS for HZNP between May 2021 and May 2024. Of 35 patients screened, 7 were excluded: 3 had disease duration >3 months (chronic PHN), 2 had incomplete follow-up data, 1 had a cardiac pacemaker, and 1 had coagulation dysfunction. The remaining 28 patients formed the study cohort, with 14 managed with inpatient programming and 14 with home-based remote programming.

Inclusion criteria were: (1) clinical diagnosis of HZNP based on standard diagnostic criteria;³ (2) age ≥18 years; (3) disease duration ≤3 months (acute or subacute phase); (4) preoperative VAS score ≥6; (5) inadequate pain relief despite standardized pharmacotherapy; and (6) complete medical records and follow-up data. Exclusion criteria were: (1) coagulation dysfunction; (2) infection at the puncture site; (3) severe cardiopulmonary, hepatic, or renal insufficiency; (4) implanted cardiac pacemaker or other electronic devices; (5) pregnancy or lactation; (6) cognitive impairment or inability to cooperate with remote programming; and (7) chronic PHN with disease duration >3 months.

All patients in both groups met the eligibility criteria for home-based programming—no cognitive impairment, no contraindication to early discharge, and a caregiver available at home. Assignment to inpatient or home programming was based on patient preference and bed availability at the time of treatment, not on clinical grounds. The two groups were therefore comparable in terms of suitability for either management strategy.

Intervention

All procedures were performed by the same experienced pain physician team. Patients were placed prone. Under C-arm fluoroscopic guidance, a 3873 Test Stimulation Lead (Medtronic Inc., Minneapolis, MN, USA) was percutaneously inserted into the epidural space at the vertebral level corresponding to the painful dermatome. The 3873 is a temporary percutaneous electrode designed specifically for short-term test stimulation; it is placed via a minimally invasive percutaneous approach under local anesthesia, requires no surgical laminectomy, and can be removed at the bedside. After intraoperative testing confirmed complete paresthesia coverage of the painful area, the lead was secured and connected through a 355531 Multi-Lead Trialing Cable (Medtronic) to an external pulse generator (Medtronic; model not recorded in all cases).

Initial stimulation parameters were frequency 40–60 Hz, pulse width 200–400 μs, with amplitude titrated to produce comfortable paresthesia covering the painful area. The electrode was retained for 10–17 days (median 14 days) and then removed at the bedside.

For the inpatient group, patients remained hospitalized throughout the stimulation period, with daily bedside programming by a physician. For the home group, patients were discharged 1–2 days after implantation with the external generator in place. Programming adjustments were made remotely by telephone or video, with the physician guiding the patient or caregiver through parameter changes. Home-group patients returned to the outpatient clinic for dressing changes every 3–5 days and for scheduled follow-up. All patients received standardized pharmacotherapy (gabapentin or pregabalin, with or without non-steroidal anti-inflammatory drugs) throughout the study period.

Outcome Measures

Patients were assessed preoperatively (T0), at electrode removal (T1), and at 3 months (T2), 6 months (T3), and 12 months (T4) postoperatively.

The primary efficacy outcome was pain intensity measured by the 11-point visual analogue scale (VAS, 0 = no pain, 10 = worst imaginable pain). A between-group difference of 2 points or more on the VAS was considered clinically meaningful, consistent with published minimal clinically important difference (MCID) estimates for neuropathic pain.

Secondary efficacy outcomes included: (1) morphine milligram equivalents (MME), calculated from all daily opioid medications, assessed at baseline and at 3 months; (2) Pittsburgh Sleep Quality Index (PSQI, 0–21, higher = poorer sleep); (3) Hospital Anxiety and Depression Scale (HADS) total score (0–42, higher = greater distress); the HADS contains anxiety and depression subscales, but subscale data were incomplete in this retrospective cohort, so only the total score is reported; (4) 36-Item Short Form Health Survey (SF-36, 0–100, higher = better quality of life). PSQI, HADS, and SF-36 were assessed preoperatively and at 3 and 12 months.

Safety outcomes comprised complications during the stimulation period and within 30 days after electrode removal: puncture-site infection, lead migration, abnormal electrical sensation, and other adverse events. Abnormal electrical sensation was defined as any unpleasant or painful stimulation-related sensation (burning, stinging, shock-like discomfort, or paresthesia outside the intended coverage area) rated ≥4 on a 0–10 scale, or any sensation the patient found intolerable and that required parameter adjustment or discontinuation of stimulation.

The long-term outcome was incidence of PHN (pain persisting >3 months after rash onset) at 12 months. PHN-free survival was estimated by the Kaplan-Meier method.

Health Economic Analysis

We adopted a limited societal perspective encompassing direct medical costs (payer perspective) and indirect costs (patient and caregiver perspective). Costs are reported in Chinese Yuan (CNY) for the treatment years 2021–2024; no inflation adjustment was applied given the relatively short study period and stable cost structure, and no discounting was applied as the time horizon did not exceed 1 year.

Direct medical costs—bed fees, nursing fees, medication costs, laboratory and imaging examination fees, and procedure fees—were extracted from the hospital information system (HIS). Electrode and consumable costs (16,000–19,000 CNY per patient, median 17,500 CNY) were identical for both groups and were excluded from the primary analysis because they do not affect the incremental comparison; a sensitivity analysis including them was performed. Indirect costs—transportation, caregiver accommodation and meals, and lost wages for patients and primary caregivers—were collected by structured patient interview at follow-up. Lost wages were valued by the human-capital approach using the average daily wage in Yibin City (approximately 200 CNY/day) multiplied by workdays lost.

An incremental cost-effectiveness ratio (ICER) was calculated using 12-month VAS score as the effectiveness measure. One-way sensitivity analysis with ±20% fluctuation in all cost items was performed to assess robustness.

Statistical Analysis

Analyses were performed with SPSS 26.0 (IBM Corp., Armonk, NY, USA) and R version 4.2.0. Normally distributed continuous variables are presented as mean ± standard deviation; between-group comparisons used independent-samples t-tests. Categorical variables are presented as n (%) and compared with Fisher’s exact test, which was used for all categorical comparisons because expected cell counts were below 5 in most cases given the small sample. Repeated VAS measurements were analyzed with repeated-measures ANOVA, reporting group main effect, time main effect, and group×time interaction effect with their respective degrees of freedom. PHN-free survival was estimated by the Kaplan-Meier method and compared with the Log rank test.

Bonferroni correction was applied for multiple comparisons: for VAS across 5 postoperative time points, the adjusted threshold was α′ = 0.01; for patient-reported outcomes across 3 instruments and 3 time points, α′ = 0.006. Both unadjusted and adjusted findings are noted in the text. Between-group mean differences with 95% confidence intervals (CIs) are reported for VAS outcomes.

No prospective sample size calculation was performed, as this was a retrospective study. Post-hoc power analysis with G*Power 3.1 showed that, with 14 patients per group and α = 0.05 (two-tailed), the study had approximately 25% power to detect a medium effect (d = 0.5) and 53% power for a large effect (d = 0.8) in between-group VAS comparisons. Approximately 64 patients per group would be needed for 80% power at d = 0.5. A two-sided P < 0.05 was considered statistically significant, except where adjusted thresholds applied. Clinical significance was assessed independently of statistical significance, using published MCID estimates where available.

Ethics Approval and Registration

This retrospective study was approved by the Ethics Committee of The First People’s Hospital of Yibin (approval no. 2025-shen-73, approved March 2025). The requirement for informed consent was waived by the ethics committee because the study involved retrospective analysis of de-identified clinical data with no more than minimal risk and no intervention beyond standard clinical care. Data analysis was initiated only after ethics approval was obtained. The study was conducted in accordance with the Declaration of Helsinki.

The study was retrospectively registered in the Chinese Clinical Trial Registry (ChiCTR2500112923, registered January 2025) and the Chinese Medical Research Registration System (MR-51-26-025284). The authors acknowledge that registration occurred after data collection was completed, which is a limitation of the study.

Results

Baseline Characteristics

Twenty-eight patients were included (14 per group). Baseline characteristics are shown in Table 1. The groups were well matched: no significant difference was found in age (63.1±6.8 vs 61.0±6.5 years, P=0.418), sex distribution (8 male/6 female vs 7 male/7 female, P=0.705), disease duration (19.9±5.6 vs 17.9±5.3 weeks, P=0.341), preoperative VAS score (8.2±0.8 vs 8.0±0.8, P=0.500), pain location (thoracic/lumbar: 9/5 vs 10/4, P=0.686), duration of electrode implantation (14.1±1.6 vs 14.6±1.2 days, P=0.360), or baseline MME (42.5±8.3 vs 40.8±7.9 mg/day, P=0.572). All patients in the cohort had pain in the thoracic or lumbar dermatomes; no patients with cervical or trigeminal distribution were included in this series.

Table 1.

Baseline Characteristics of the Two Groups

Characteristic	Inpatient (n=14)	Home-Based (n=14)	P value
Age (years), mean±SD	63.1±6.8	61.0±6.5	0.418
Sex (male/female), n	8/6	7/7	0.705
Disease duration (weeks), mean±SD	19.9±5.6	17.9±5.3	0.341
Preoperative VAS, mean±SD	8.2±0.8	8.0±0.8	0.500
Pain location (thoracic/lumbar), n	9/5	10/4	0.686
Electrode implantation duration (days), mean±SD	14.1±1.6	14.6±1.2	0.360
Baseline MME (mg/day), mean±SD	42.5±8.3	40.8±7.9	0.572

Notes: Continuous variables compared with independent-samples t-test; categorical variables with Fisher’s exact test. All patients had pain in the thoracic or lumbar dermatomes.

Abbreviations: VAS, visual analogue scale; MME, morphine milligram equivalents.

Pain Intensity (VAS)

VAS scores at each time point are presented in Table 2. Repeated-measures ANOVA showed a strong effect of time (F(4, 104)=632.78, P<0.001), confirming that both groups experienced substantial and sustained pain reduction after st-SCS. The group main effect was not significant (F(1, 26)=1.008, P=0.325), and neither was the group×time interaction (F(4, 104)=0.261, P=0.902), indicating that the trajectory of pain relief was similar between the two management strategies.

Table 2.

VAS Scores at Each Time Point

Time Point	Inpatient (n=14)	Home-Based (n=14)	Mean Difference (95% CI)	P value
Preoperative	8.2±0.8	8.0±0.8	0.2 (−0.38, 0.78)	0.500
At electrode removal	2.3±0.6	2.1±0.5	0.2 (−0.21, 0.61)	0.348
3 months	1.4±0.6	1.1±0.5	0.3 (−0.13, 0.73)	0.178
6 months	1.3±0.6	1.0±0.5	0.3 (−0.12, 0.72)	0.159
12 months	1.6±0.8	1.1±0.7	0.5 (−0.07, 1.07)	0.087
MME at 3 months (mg/day)	18.0±4.2	13.2±4.0	4.8 (1.64, 7.96)	0.005

Notes: Data are mean±SD unless otherwise indicated. P-values from independent-samples t-tests. Repeated-measures ANOVA: time effect F(4, 104)=632.78, P<0.001; group effect F(1, 26)=1.008, P=0.325; group×time interaction F(4, 104)=0.261, P=0.902. All between-group VAS differences are below the minimal clinically important difference of approximately 2 points.

Abbreviations: VAS, visual analogue scale; MME, morphine milligram equivalents.

More relevant from a clinical standpoint, the between-group difference in VAS scores was less than 0.5 points at every postoperative time point: at electrode removal (2.3±0.6 vs 2.1±0.5, mean difference 0.2, 95% CI −0.21 to 0.61, P=0.348), at 3 months (1.4±0.6 vs 1.1±0.5, mean difference 0.3, 95% CI −0.13 to 0.73, P=0.178), at 6 months (1.3±0.6 vs 1.0±0.5, mean difference 0.3, 95% CI −0.12 to 0.72, P=0.159), and at 12 months (1.6±0.8 vs 1.1±0.7, mean difference 0.5, 95% CI −0.07 to 1.07, P=0.087). These differences are well below the minimal clinically important difference of approximately 2 points for neuropathic pain, meaning that even the upper bounds of the confidence intervals do not approach a clinically meaningful gap. In practical terms, patients in both groups obtained comparable pain relief regardless of whether programming was done in hospital or at home.

At 3 months, the home group had lower daily opioid consumption than the inpatient group (13.2±4.0 vs 18.0±4.2 mg/day, mean difference 4.8 mg, 95% CI 1.64 to 7.96, P=0.005). A difference of roughly 5 mg morphine equivalents per day translates to about one extra tablet of combination opioid analgesic daily; whether this is clinically meaningful for an individual patient depends on the specific medication and context, but at a group level it indicates lower opioid exposure in the home-managed patients.

Patient-Reported Outcomes

Patient-reported outcomes are presented in Table 3. Preoperatively, the groups did not differ in PSQI (11.2±2.1 vs 10.8±1.9, P=0.581), HADS total score (12.5±2.8 vs 11.8±2.5, P=0.481), or SF-36 (52.3±8.5 vs 54.1±7.9, P=0.570).

Table 3.

Patient-Reported Outcomes

Outcome	Time Point	Inpatient (n=14)	Home-Based (n=14)	P value
PSQI	Preoperative	11.2±2.1	10.8±1.9	0.581
	3 months	7.1±1.5	5.8±1.3	0.023*
	12 months	6.3±1.4	5.5±1.2	0.115
HADS total	Preoperative	12.5±2.8	11.8±2.5	0.481
	3 months	7.8±2.1	6.2±1.8	0.041*
	12 months	6.5±1.9	5.6±1.6	0.176
SF-36	Preoperative	52.3±8.5	54.1±7.9	0.570
	3 months	72.5±6.8	78.4±6.2	0.023*
	12 months	76.8±6.5	80.2±5.8	0.153

Notes: Data are mean±SD. P-values from independent-samples t-tests. *P<0.05 unadjusted; not significant after Bonferroni correction for 9 comparisons (adjusted α′=0.006). HADS subscale data were incomplete and are not reported.

Abbreviations: PSQI, Pittsburgh Sleep Quality Index; HADS, Hospital Anxiety and Depression Scale; SF-36, 36-Item Short Form Health Survey.

At 3 months, the home group had numerically lower PSQI (5.8±1.3 vs 7.1±1.5, P=0.023) and HADS total scores (6.2±1.8 vs 7.8±2.1, P=0.041) and higher SF-36 scores (78.4±6.2 vs 72.5±6.8, P=0.023). However, after Bonferroni correction for multiple comparisons across 3 instruments and 3 time points (adjusted α′=0.006), none of these differences retained statistical significance. From a clinical perspective, the PSQI difference of 1.3 points is below the commonly cited MCID of 3 points for the PSQI; the HADS difference of 1.6 points sits at the lower edge of what might be considered clinically noticeable; and the SF-36 difference of 5.9 points falls within the range often regarded as a small but potentially meaningful improvement. Taken together, these findings suggest a possible early advantage for home management in psychological well-being and quality of life, but the evidence is not strong enough to draw firm conclusions. At 12 months, no significant differences were found in any patient-reported outcome (all P>0.05), and the between-group differences had narrowed.

Safety and Complications

Complications are presented in Table 4. The overall complication rate was 14.3% (2/14) in both groups (Fisher’s exact test, P=1.000). In the inpatient group, one patient developed a puncture-site infection that resolved with oral antibiotics and local wound care, and one experienced abnormal electrical sensation that resolved after parameter adjustment. In the home group, one patient had lead migration that required repositioning during an outpatient visit, and one had abnormal electrical sensation resolved by remote parameter adjustment. No serious adverse events, epidural hematoma, or neurological deficits occurred in either group, and no patient required early electrode removal because of a complication. The complication types differed in a way that is clinically intuitive: the inpatient group had an infection (prolonged bed rest and hospital exposure), while the home group had lead migration (patient mobility after early discharge). Both were managed successfully.

Table 4.

Complications, PHN, and Healthcare Costs

Outcome	Inpatient (n=14)	Home-Based (n=14)	P value
Complications, n (%)	2 (14.3)	2 (14.3)	1.000
Puncture site infection	1	0	1.000
Lead migration	0	1	1.000
Abnormal electrical sensation	1	1	1.000
PHN at 12 months, n (%)	2 (14.3)	1 (7.1)	1.000
Direct medical cost (CNY), mean±SD	10,857.1±1025.3	8428.6±760.4	<0.001
Indirect cost (CNY), mean±SD	2378.6±523.5	1035.7±243.1	<0.001
Total cost (CNY), mean±SD	13,235.7±1498.6	9464.3±937.5	<0.001
Cost reduction (%)	–	28.5%	–

Notes: Categorical variables compared with Fisher’s exact test; continuous variables with independent-samples t-tests. Costs exclude electrode and consumable fees (16,000–19,000 CNY per patient, identical for both groups). Complication subtypes are not mutually exclusive; each patient may have had only one complication.

Abbreviations: PHN, postherpetic neuralgia; CNY, Chinese Yuan.

Postherpetic Neuralgia (PHN)

At 12 months, PHN had developed in 2 of 14 patients (14.3%) in the inpatient group and 1 of 14 (7.1%) in the home group (Fisher’s exact test, P=1.000). With only three events total, the study had essentially no ability to detect a difference in PHN incidence. Kaplan-Meier analysis of PHN-free survival showed no significant difference between groups (Log rank test: χ²=0.359, P=0.549; Figure 1). The single-patient difference in PHN rates is not interpretable as a treatment effect and should not be used to infer a mechanism.

Figure 1.

Kaplan-Meier curves for PHN-free survival after short-term spinal cord stimulation. The blue line represents the inpatient programming group (n=14), and the red line represents the home-based programming group (n=14). Numbers at risk are shown below the x-axis. Log rank test: χ²=0.359, P=0.549. With only three PHN events total (two in the inpatient group, one in the home group), this analysis is descriptive and underpowered.

Abbreviation: PHN, postherpetic neuralgia.

Health Economic Analysis

Health economic outcomes are presented in Table 4 and Table 5. The total healthcare cost was lower in the home group (9464.3±937.5 vs 13,235.7±1498.6 CNY, P<0.001), a reduction of 3771 CNY per patient, or 28.5%. This cost difference is not attributable to differences in analgesic efficacy—VAS scores were comparable between groups—but rather to the shorter hospital stay in the home group. The cost breakdown (Table 5) makes this clear: bed fees were 56.5% lower in the home group (1850 vs 4250 CNY) and nursing fees were 63.1% lower (620 vs 1680 CNY), while medication costs were actually slightly higher in the home group (3120 vs 2850 CNY), as would be expected when patients manage their medications at home rather than receiving them through inpatient pharmacy.

Table 5.

Detailed Cost Component Breakdown (CNY)

Cost Component	Inpatient (n=14)	Home-Based (n=14)
Direct medical costs
Bed fees	4250.0±580.2	1850.0±320.5
Nursing fees	1680.0±220.3	620.0±110.2
Medication costs	2850.0±380.5	3120.0±410.3
Examination fees	1250.0±180.2	1480.0±200.5
Procedure fees	827.1±120.5	1358.6±180.2
Indirect costs
Transportation	680.0±120.3	320.0±60.5
Caregiver expenses	1050.0±220.5	420.0±80.3
Lost wages	648.6±150.2	295.7±70.5

Notes: Data are mean±SD. Costs exclude electrode and consumable fees (16,000–19,000 CNY per patient, identical for both groups). Procedure fees include electrode implantation and removal fees. Direct medical costs extracted from hospital information system (HIS); indirect costs collected by structured patient interview. Lost wages valued by human-capital approach using average daily wage in Yibin City (~200 CNY/day).

Direct medical costs were lower in the home group (8428.6±760.4 vs 10,857.1±1025.3 CNY, P<0.001), as were indirect costs (1035.7±243.1 vs 2378.6±523.5 CNY, P<0.001). The indirect cost savings reflect reduced transportation, caregiver accommodation, and lost wages associated with shorter hospital stays. When electrode and consumable costs (median 17,500 CNY per patient) were included in a sensitivity analysis, the total cost remained lower in the home group (26,964.3±937.5 vs 30,735.7±1498.6 CNY, P<0.001), though the percentage reduction narrowed to 12.3% because the fixed device cost dominated the total. The ICER was −7543 CNY per additional 1-point VAS reduction at 12 months, indicating that home programming was less costly and was associated with numerically (though not significantly) lower VAS scores. However, because the VAS difference was below the MCID, the ICER should be interpreted as reflecting cost savings from shorter hospitalization rather than a true gain in cost-effectiveness per unit of pain relief. One-way sensitivity analysis with ±20% fluctuation in all cost items confirmed that the cost savings were robust (all P<0.001).

Discussion

In this retrospective cohort of 28 patients with acute or subacute HZNP treated with short-term spinal cord stimulation, we found that managing patients at home with remote programming produced pain relief comparable to that achieved with inpatient programming. The word “comparable” here is not a euphemism for “we failed to find a difference”; the between-group gap in VAS scores was under half a point at every time point, well below the two-point threshold generally regarded as clinically meaningful for neuropathic pain. In other words, even if a larger study rendered these differences statistically significant, they would not matter to a patient. This is the central clinical finding of the study, and it held up under repeated-measures ANOVA, which showed no group effect and no group-by-time interaction.

The pain relief itself was substantial in both groups—VAS scores fell from roughly 8 points preoperatively to 1–2 points after stimulation and remained low through 12 months.^7–9 The durability of SCS analgesia beyond 12 months has been demonstrated in longer-term follow-up studies.¹⁰ This is consistent with the growing body of literature supporting st-SCS as an effective intervention^3,11–13 for acute and subacute HZNP, and it reinforces the notion that the acute/subacute window is a favorable time to intervene. What our study adds is the observation that where the programming happens—in hospital or at home—does not appear to determine how much pain relief the patient gets. The stimulation parameters, the electrode placement, and the patient’s underlying condition matter far more than the setting in which parameters are adjusted.^3,11–13

Patient-reported outcomes told a more nuanced story. At 3 months, the home group had lower PSQI and HADS scores and higher SF-36 scores, but these differences did not survive Bonferroni correction, and by 12 months they had largely disappeared. We are hesitant to make much of the 3-month findings. The PSQI difference of 1.3 points is below what most clinicians would consider a meaningful improvement in sleep quality. The HADS difference of 1.6 points is at the lower edge of clinical relevance. The SF-36 difference of nearly 6 points is more suggestive, but in the context of multiple uncorrected comparisons it could well be a chance finding. What seems plausible, even if not proven, is that recovering at home is easier on a patient’s mood and sleep than recovering in a hospital ward. Hospital stays disrupt routines, expose patients to noise and anxiety, and separate them from family. It would be surprising if that did not affect psychological well-being, at least early on. But the effect appears transient, and we would not want to overstate it.

The lower opioid consumption in the home group at 3 months (13.2 vs 18.0 mg/day MME) is worth noting. A difference of about 5 mg morphine equivalents is roughly one extra combination opioid tablet per day—not dramatic for an individual, but meaningful at a population level, where lower opioid exposure translates into fewer adverse effects and lower risk of dependence. The mechanism is not clear from our data. It may reflect the better psychological state of home-managed patients, since anxiety and depression are known to amplify pain and drive opioid use.¹⁴ Or it may simply be that patients at home are less likely to request additional analgesia than patients on a hospital ward, where nursing staff routinely offer pain medication. We cannot distinguish these possibilities with a retrospective design.

Safety was reassuring. The complication rate was 14.3% in both groups, with no serious adverse events in either. The types of complications differed in an expected way: the inpatient group had a puncture-site infection (prolonged bed rest and hospital exposure), while the home group had a case of lead migration (patient mobility after early discharge). Both were managed without difficulty. The 14.3% rate sits within the 5–20% range reported for temporary SCS^15–17 in the literature. For clinicians considering home management, the takeaway is that externalized leads can be managed safely outside the hospital, provided patients are carefully selected, given clear instructions, and seen at regular intervals for dressing changes and lead checks. The lead migration case in our series was picked up at a scheduled outpatient visit and corrected the same day; close follow-up is therefore not optional.

The cost analysis is where the two strategies diverge most clearly. Home management cost 28.5% less per patient, a saving of about 3770 CNY. This is not because home patients got better pain relief—they did not—but because they spent fewer days in hospital. The cost breakdown tells the story: bed fees and nursing fees were roughly 60% lower in the home group, while medication costs were actually slightly higher, as one would expect when patients self-administer medications at home rather than receiving them through an inpatient pharmacy. Even after including the substantial fixed cost of the electrode and consumables (16,000–19,000 CNY), the home group remained 12.3% cheaper. For a hospital system, the implication is straightforward: discharging st-SCS patients early with remote programming frees beds, reduces nursing workload, and lowers costs without compromising pain control.^4,18 The magnitude of savings will vary by healthcare setting—in systems where bed fees are lower or where telemedicine infrastructure requires additional investment, the gap may narrow.^18,19 But the direction of the effect is unlikely to reverse.²⁰

Several limitations need to be acknowledged. The sample size of 28 patients is small, and the post-hoc power analysis showed only 25% power to detect a medium effect. This means that while the VAS differences we observed are too small to be clinically meaningful regardless of sample size, we cannot rule out the possibility that some other outcome—PHN incidence, for instance—might differ between groups in a way our study was too small to detect. The retrospective design and non-randomized assignment (based on patient preference and bed availability) introduce potential selection bias, although baseline characteristics were well balanced. The single-center setting limits generalizability, particularly to healthcare systems with different cost structures or telemedicine reimbursement models. Indirect cost estimates relied on patient self-report and average wage data, which introduces measurement error. We did not assess patient satisfaction with remote programming, an outcome that matters to patients and that may influence willingness to adopt this approach. The HADS subscale data were incomplete, preventing analysis of anxiety and depression separately. Finally, the study was registered retrospectively, after data collection was completed, and ethics approval was obtained after the clinical care was delivered; while the data analysis was not initiated until after approval, the timeline is not ideal and reflects the practical realities of conducting retrospective research in a busy clinical department.

Despite these limitations, the study provides a practical signal for clinicians managing HZNP with st-SCS: with appropriate patient selection and follow-up, home-based remote programming is a viable alternative to inpatient management that saves money and may improve early psychological well-being, without sacrificing pain control. Larger, multicenter, prospective studies—ideally randomized—are needed to confirm these findings and to identify which patients benefit most from remote management. Future work should also include patient-reported satisfaction, longer follow-up, and a more granular analysis of complication types and their management in the home setting.

Conclusion

For patients with acute or subacute HZNP treated with short-term spinal cord stimulation, early discharge with remote programming produced pain relief comparable to inpatient management—comparable not merely in the statistical sense of failing to detect a difference, but in the clinical sense that the between-group gap was well below the threshold of meaningful pain relief. Home management was associated with lower early opioid consumption, numerically better psychological and quality-of-life outcomes at 3 months (though not after correction for multiple comparisons), similar complication rates, and substantially lower healthcare costs. The cost saving reflects shorter hospital stays, not superior analgesia. These findings support remote programming as a practical alternative for appropriately selected patients, but they should be confirmed in larger prospective studies before being adopted as standard practice.

Funding Statement

This study was supported by the Hospital-level Research Fund of The First People’s Hospital of Yibin (grant no. 2025-KYY-9). The funding body had no role in the design of the study, collection, analysis, and interpretation of data, or in writing the manuscript.

Data Sharing Statement

The de-identified individual participant data generated and analyzed in this study, along with the statistical analysis code, are available from the corresponding author (Shiming Huang, 396942192@qq.com) on reasonable request, subject to institutional data sharing policies and ethical review. Data will be made available after publication and for at least 5 years thereafter. Requests should include a brief description of the intended use and will be reviewed within 2 weeks.

Ethics Approval and Consent to Participate

This retrospective study was approved by the Ethics Committee of The First People’s Hospital of Yibin (approval no. 2025-shen-73, approved March 2025). The requirement for informed consent was waived by the ethics committee because the study involved retrospective analysis of de-identified clinical data with no more than minimal risk and no intervention beyond standard clinical care. Data analysis was initiated only after ethics approval was obtained. The study was conducted in accordance with the Declaration of Helsinki.

Consent for Publication

Consent for publication was not applicable.

Author Contributions

All authors made a significant contribution to the work reported, whether that is in the conception, study design, execution, acquisition of data, analysis and interpretation, or in all these areas; took part in drafting, revising or critically reviewing the article; gave final approval of the version to be published; have agreed on the journal to which the article has been submitted; and agree to be accountable for all aspects of the work.

Disclosure

The authors declare that they have no competing interests.