Contemporary Landscape of Medical Technology in Gastroenterology Between 2013 and 2023

Abstract

Medical technology in gastroenterology spans devices, point-of-care diagnostics, digital health and artificial intelligence. We conducted a narrative review of US Food and Drug Administration-reviewed medical technology on the available evidence base to support uptake in clinical practice. 713 new GI devices were FDA-approved or cleared between 2013 and 2023. The most frequent technologies had indications for use in general endoscopy (442 or 62.0% of all new technologies) with advanced endoscopy technologies being the second most common (172 or 24.1%). Ninety-nine new technologies had indications for use outside of the endoscopy suite. 99% of new technologies were FDA-cleared against existing technologies via the 510(k) pathway rather than approved on new clinical data. Among the 19 devices with clinical trial data, trials were limited in sample size and generally evaluated disease-related changes or technical success rather than patient-reported outcomes that anchor drug trials. Twelve devices (63.2%) had favorable, published cost-effective data. To promote evidence-based discussion and uptake of new technology, and especially disruptive or high-risk technology, we propose a practical framework informed by balancing probable efficacy and safety and considering cost-effectiveness.

INTRODUCTION

The scope of medical technology in gastroenterology

Medical technology is essential to gastroenterology, with colonoscopy and endoscopy being among the most performed procedures. Over the past two decades, interventional endoscopy has become a recognized subspecialty spanning a clinical footprint across bariatrics, gastrointestinal luminal disease, and hepatobiliary/pancreatic disease. Outside of the endoscopy suite, the usage of digital tools and artificial intelligence has been emerging across many gastroenterology subspecialities including inflammatory bowel disease, hepatology, and motility.

Limited awareness of med tech regulation

Despite daily reliance on medical technology, >80% of physicians (including gastroenterologists) remain largely unaware that FDA’s required scientific data for medical technology differs from the expected scientific data for drugs.(1) For example, the most recent colonoscopes used for colon cancer screening in every day practice were FDA cleared without requiring clinical trial data.(2-4) While the colonoscope was cleared by the FDA via the 510(k) pathway as it was compared to a precedent device, this reality stands at odds with the regulations required for drug approval, as every new drug is evaluated based on systematic reviews of large clinical trials to set the standard for shared decision-making, evidence-based medicine, and practice guideline development. Despite medical technology being integral to gastroenterology, little is known about the FDA medical device approval/clearance pathways and the clinical data required.

Med tech regulatory pathways

There are five main pathways for FDA approval/clearance of medical devices. Each new medical device is assigned one of three regulatory classes: Class I, II, and III. As devices increase in class, the level of regulatory control also increases.(5) Most Class I devices and some of Class II devices can become exempt from the 510(k) pathway if the FDA determines that the device is safe and effective.(6) A 510(k) premarket submission is used for Class I and Class II devices that have a substantially equivalent legally marketed device.(7) A Premarket Approval (PMA) is required for Class III devices, which are defined as devices that can support or sustain human life but can pose potential and unreasonable risk or injury.(8) De Novo devices are novel medical devices that have never been marketed in the US. They are compared to controls, and the use of these controls can provide reasonable assurance of safety and efficacy. These devices are also classified into regulatory classes based on their risks.(9) Lastly is the Humanitarian Device Exception pathway, which are all Class III devices intended to be used in patients with rare diseases or conditions (<4,000 patients per year).(10)

Purpose of this review

We performed a review to synthesize and summarize the available evidence on medical technology development in gastroenterology. However, we recognize that many factors influence the utilization of new medical devices in the clinical practice, such as device cost, device efficacy, and even clinician exposure and familiarity. Additionally, we reviewed the cost-effectiveness of these new devices with supported clinical data. Our goal is to provide physicians with further insights into the approval processes of new medical devices and to assess whether the cost of the device is justified by its effectiveness. We aim to empower physicians to make practical recommendations that support benefit and risk discussions surrounding utilization of medical technology in health care, as well as inform them of the potential economic benefits associated with each device utilization.

METHODS

Search Strategy and Eligibility Criteria

Our research librarian team developed our search strategy, and included FDA databases for 510(k) clearance, De Novo classification, and pre-market approval (PMA) to identify medical device applications that were FDA-cleared or approved with decision dates between 2013 and 2023.^11-13 Search terms included limiting results to the Gastroenterology/Urology Panel or the Gastroenterology/Urology Advisory Committee. Results were extracted to Excel 365 (Microsoft Corporation, Redmond, WA). We excluded devices receiving Humanitarian Device Exemption and those indicated for use outside gastroenterology and hepatology.

Review

Three research team members (MH, SX, RS) independently and manually reviewed all device applications, and discrepancies in eligibility were resolved by consensus with the senior investigator (ES). Applications related to the same device were consolidated for reporting purposes.

Data Extraction and Analysis

Devices were classified as general endoscopy (including endoscopic devices or accessories), advanced endoscopy (including devices used for advanced endoscopic procedures), and non-endoscopic gastroenterology devices (including devices used for motility, nutrition, or transplant). Extracted information for 510(k) devices included device name, 510(k) or PMA or de novo registration number, decision date, presence of FDA-reviewed clinical data in the approval letter, and predicate device (when applicable).

Clinical trials used to support FDA clearance or approval were cross-linked using PubMed and EMBASE directly from FDA databases to complete data extraction. For applications containing clinical data, we extracted data from published clinical trials to assess the quality of evidence in the study conducted according to the CONSORT statement, which reflects the gold standard for drug trial reporting. These data included study type, trial length, patient population, recruitment strategy, number of patients, randomization method (if used), comparator (if present), study outcomes/endpoints, and statistical plan.

To assess risk profiles for new gastroenterology technologies, the number and types of adverse and serious adverse events were extracted as reported by the study authors.

To assess the cost-effectiveness of the devices with supporting clinical trials, we conducted a literature search using PubMed. The search utilized keywords such as the proprietary and non-proprietary names of each device, "cost-effectiveness analysis," and "cost-benefit analysis." For each relevant analysis, we reviewed the abstract and full article to extract data, including whether the study adhered to the PRISMA-ScR checklist, cost of the device, patient population, time horizon, primary economic savings, and primary health outcome.

THE MED TECH PIPELINE

Seven-hundred and thirteen new gastroenterology technologies received FDA approval or clearance between 2013 and 2023.(11-13) Of this total, seven were FDA-approved via the PMA pathway and fifteen were approved via the De Novo pathway. The other 691 new technologies were FDA-cleared through the 510(k) pathway as substantially equivalent to an existing device.

Technology development across sub-specialties of gastroenterology

More than half of the technology via the 510(k) pathway (433 or 62.7%) and the De Novo pathway (8 or 53.3%) were intended for use in general endoscopy. In the PMA pathway, almost all devices are intended to be used outside of the endoscopy suite (6 or 85.7%) (Table 1). Across gastroenterology subspecialties, almost one-half of technologies were intended for use in general gastroenterology (347 or 48.7% of all new technologies), with endoscopic accessory devices being second-most common (117 or 16.4% of all new technologies) and hepatic/hepatobiliary technologies being third most common (105 or 14.7% of all new technologies).

Table 1:

Number of new FDA-approved or FDA-cleared technologies in gastroenterology between 2013 and 2013, stratified by sub-specialty and availability of clinical data evaluating intended use.

Setting for intended use	510(k)		Pre-Market Approval		De Novo
	Number of Technologies	Technologies with Clinical Data	Number of Technologies	Technologies with Clinical Data	Number of Technologies	Technologies with Clinical Data
Advanced Endoscopy	169	1	0	0	3	3
General Endoscopy	433	6	1	1	8	5
Non-Endoscopic Gastroenterology	89	1	6	6	4	4
Sub-specialty of intended use	Number of Technologies	Technologies with Clinical Data	Number of Technologies	Technologies with Clinical Data	Number of Technologies	Technologies with Clinical Data
General gastroenterology	333	6	4	4	9	6
Advanced Endoscopy	40	0	0	0	0	0
Endoscopy Accessory	117	0	0	0	0	0
Esophageal	28	0	1	1	1	1
Hepatic/Hepatobiliary	103	1	0	0	2	2
Motility	20	1	0	0	1	1
Nutrition	36	0	0	0	0	0
Pancreas	14	0	0	0	2	2
Transplant	0	0	2	2	0	0
Total	691	8	7	7	15	12

Availability of clinical evidence

Clinical trial data included in an application to the FDA are rigorously evaluated during the regulatory process because these data can be used in FDA decision-making and in marketing claims for FDA-approved or FDA-cleared products. Eight of the 691 510(k) devices that came to market between 2013 and 2023 had clinical data supporting intended use, representing 1.1% of all new 510(k) devices. 98.9% of all new 510(k) devices did not include clinical data in their FDA application. Among devices that are first-of-its-kind for intended use, all PMA devices (including the highest level of risk to patients) and most de novo devices (low or moderate risk) included clinical data to evaluate their intended use.

Quality of clinical evidence

Among the 27 clinical trials conducted between 2013 and 2023 to support FDA approval or clearance of new, non-drug technologies in gastroenterology, 18.5% were randomized controlled trials of adequate sample size to detect statistical change on a primary endpoint.(14-18) Apart from four motility devices(19-22), only one other trial(23) used patient-reported outcomes as a primary endpoint (Supplemental Table 1 for a description of clinical trials). Instead, most trials targeted objective endpoints such as objective disease improvement or technical success of device deployment (such as biliary stent placement, lumen opposing stents for pancreatic pseudocysts, or direct endoscopic necrosectomy).

Safety

Among the 1.1% of 510(k) devices with clinical data accompanying their FDA application for clearance as being substantially equivalent to existing technology, seven of the eight device trials reported no adverse events.(15,16,23-27) Minor and transient adverse events with one 510(k) device intended for hepatobiliary use for stenting of malignant obstructions (percutaneous intraductal RFA followed by dual stent placement) that received 510(k) clearance.(28)

Adverse events were common in clinical trials evaluating first-in-class technologies and intended-use applications. For example, adverse events occurred in eight (72.7%) of the 11 clinical trials for de novo devices deemed to have a low or moderate risk to patients, and serious adverse events occurred in two of these trials for devices intended for use in advanced endoscopy. One occurred in a trial device intended for use in general endoscopy. These serious adverse events represented known complications of endoscopic pancreatic pseudocyst drainage, colonoscopy, and stenting of biliary strictures, leading these devices to ultimately receive de novo classification.(29,30) All trials that supported pre-market approval (i.e., devices with the highest level of risk to patients) of new technologies reported serious adverse events in clinical trials. No device-related deaths were reported in clinical trials that supported any FDA approval or clearance (Tables 2 and 3 for description of adverse events in gastroenterology medical technology clinical trials over the past decade).

TABLE 2:

Overview of adverse events (AE) and serious adverse events (AE) across new gastrointestinal technologies in the United States between 2013-2023 for the De Novo pathway, Pre-Market approval pathway, and 510(k) pathway.

De novo technologies (low or medium risk to patients)
Device	Total number of patients in device trials (N)	Number of patients experiencing at least 1 AE	Total number of AEs during trials	Total number of SAEs during trials	Deaths Related to Device	Deaths Unrelated to Device
Axios Stent (Endoscopic drainage of pancreatic pseudocyst)	33	16 (48.5% of total)	45	25	0	0
Pillcam Colon 2 (Colon capsule imaging system)	884	101 (11.4%)	143	1	0	0
Eclipse (Rectal control system for fecal incontinence)	61	61 (100.0%)	93	0	0	0
Wallflex Biliary RX (Metallic stent for benign biliary strictures)	127	53 (41.7%)	--	86	0	10
IB-Stim (Non-implanted nerve stimulator for functional abdominal pain)	115	10 (8.7%)	10	0	0	0
Echotip Insight (Endoscopic transhepatic venous access needle)	28	0 (0.0%)	0	0	0	0
Mucosal Integrity Conductivity (MI) Test (Esophageal tissue characterization system)	112	1 (0.9%)	1	0	0	0
NaviCam Capsule Endoscope System (Magnetically maneuvered capsule endoscopy system)	353	5 (1.4%)	9	0	0	0
GI Genius (GI lesion software detection system)	700	0 (0.0%)	--	0	0	0
Spyglass DS Direct Visualization System (Platform to visualize and treat biliary disease)
AccuMeasure System (Endoscopic light-projecting measuring device)	Redacted	0 (0.0%)	0	0	0	0
Vibrant Ltd. (Orally ingested device for constipation)	349	62 (17.8%)	74	0	0	0
Pre-market approval technologies (includes devices with high risk to patients)
Device	Total number of patients in device trials (N)	Number of patients experiencing at least 1 AE	Total number of AEs during trials	Total number of SAEs during trials	Deaths Related to Device	Deaths Unrelated to Device
Acticon Neosphincter (Implantable device for treatment fecal incontinence)	112	102 (91.1% of total)	395	Unknown	0	0
Axonics Sacral Neuromodulation (SNM) System (Implanted electrical device for treatment of fecal incontinence)	129	80 (62.0%)	181	7	0	1
Interstim (Implanted electrical device for treatment of fecal incontinence)	120	88 (73.3%)	237	20	0	0
LINX (Implanted device for the treatment of gastroesophageal reflux disease)	100	76 (76.0%)	162	12	0	0
Organ Care System (OCS®) Liver (Perfusion system for the preservation of donor livers)	299	82 (27.4%)	Unknown	150	0	0
OrganOx metra System (Perfusion system for the preservation of donor livers)	266	95 (35.7%)	Unknown	275	0	0
SOLESTA (Injectable bulking agent for the treatment of fecal incontinence)	206	97 (47.1%)	238	2	0	0
510(k) technologies (substantially equivalent)
Device	Total number of patients in device trials (N)	Number of patients experiencing at least 1 AE	Total number of AEs during trials	Total number of SAEs during trials	Deaths Related to Device	Deaths Unrelated to Device
Elra Electrode (Endoscopic bipolar radiofrequency ablation electrode used to ablate tissue in the common bile duct)	30	3 (10.0% of total)	3	0	0	0

Note: Twelve devices reported clinical data in their FDA submissions. Data for one device (Endorotor) was not made available in the public domain.

TABLE 3:

Overview of cost-effective analyses for new medical devices approved or cleared with clinical trial data between 2013-2023.

Device	Non-Proprietary name	Study	Use of PRISMA	Patients Criteria	Time Horizon	Cost of Device	Economic Savings	Health Savings
Axios Stent or Wallflex Biliary Rx	Lumen-apposing metal stent; metal biliary stent	Chen et al (2018)(42)	Yes	Endoscopic ultrasound-guided drainage of pancreatic walled-off necrosis	6 months	$20,029 for metal stent	More costly than plastic stents; cost per successful drainage was $21,712 for metal stent and $18,990 for plastic stent. Incremental cost-effectiveness ratio: 49214 to perform 1 additional successful drainage in comparison to plastic stent.	Success rate of 92.2% with metal stent versus 83.9% with plastic stent.
		Chen et al (2018)(43)	No	Endoscopic ultrasound management of pancreatic pseudocyst	6 months	Unknown	Cost per successful drainage: $18,129 with metal stent and $10,403 with plastic stent.	Success rate was 93.9 % for metal stent and 96.96 % for plastic stent.
		Roberts et al (2021)(44)	No	Preoperative biliary drainage prior to a pancreatoduodenectomy	Unknown	$2900 AUD for metal stent	The predicted cost of using PS and self-expandable metal stent were similar [$4529 (Australian) plastic stent vs. $4408 (Australian) self-expanding metal stent].	The overall rate of complications was lower in the metal stent group compared to plastic stent group (12.2% vs. 28.7%, p = 0.024).
		Elshimi et al (2020)(45)	No	All patients ≥18 years with obstructive jaundice caused by hepatocellular carcinoma	1 year or until re-obstruction or death	Total cost for an ERCP procedure with metal stent: 14,000 EGP	Total direct and indirect costs were noted to be 41,745.84£ per patient in the plastic stent group and 22,256.62£ per patient in the metal stent group (with a difference of 19,489.24£).	Stent dysfunction and/or re-obstruction is more common with plastic stents, which can incur indirect costs with additional procedures, hospitalizations, and time lost.
		Farani et al (2021)(34)	No	Patients with malignant obstructive jaundice who underwent palliative biliary stent insertion	6 months	Median initial cost IDR 93 998 886 in the metal stent group	More costly initially: 60 663 260 Indonesian Rupiah in the plastic stent group, and 93 998 886 Indonesian Rupiah in the metal stent group. Using a decision tree model, the average total cost in 6 months with stent patent outcome for the two groups were 105 571 028 Indonesian Rupiah (proportion of patent stent was 67%) in the plastic stent group, and 104 446 233 Indonesian Rupiah (proportion of patent stent was 81%) in the metal stent group.	Mortality rate in the first month: 25.2% in plastic stent group and 30.3% in metal stent group. Mortality rate at 6-month follow-up: 64.1% in plastic stent group and 66.7% metal stent group. 30-day dysfunction rate after stent insertion: 13% in the plastic stent group and 12.1% in the metal stent group. Dysfunction rate at 180 days: 24.4% plastic stents and 15.2% metal stents.
		Martinez et al (2017)(46)	No	Adult patients in the USA with nonresectable, locally advanced (stage III/IV) pancreatic carcinoma	Unknown	Unknown	Model projected patients with initial placement of metal stents saved $1453 in total costs over a lifetime ($304,151 vs. $305,605) when compared with patients who received plastic stents.	Initial use of metal stents resulted in 0.3 months greater quality-adjusted life months (12.3 vs. 12.0 months). Initial use of metal biliary stents reduced the number of patients requiring a subsequent stent by 61% (28 vs. 89%), as well as resulted reductions in number of stents placed per patient over a patient’s lifetime (1.4 vs. 2.8).
		Gardner et al (2016)(47)	No	Patients with extrahepatic malignant biliary obstruction from newly diagnosed pancreatic adenocarcinoma	Unknown	Fully covered self-expandable metal stent (fcSEMS): $4,274 Uncovered self-expandable metal stent (uSEMS): $3,446	Stent placement cost more during initial procedure (uncovered = $24,874 and fully covered = $22,729 vs. plastic = $18,701), and they also resulted in higher procedural adverse event costs per patient (uncovered = $5521 and fully covered = $12,701 vs plastic = $0). Plastic stents resulted in $11,458 hospitalization cost per patient due to stent occlusion (compared to $2301 for uncovered and $0 for fully covered). The total cost of the index procedure, procedural adverse events associated with the index procedure, and adverse events from stent occlusion were similar between stent types (fully covered = $41,112; uncovered = $41,475; and plastic = $39,955).	Fully-covered had a longer time to stent occlusion compared with uncovered and plastic stents. Fully covered resulted in fewer total days (3) of neoadjuvant treatment delay than uncovered (39) or plastic stents (79).
		Barkun et al (2015)(48)	No	Patients with extrahepatic biliary obstruction by any malignant process	12 months	Metal stent: $6,701	On average, it would save US$13,971 per patient to opt for the initial metal stent strategy. Incremental cost-effectiveness ratio: −US$26,993/dominated for initial polyethylene stent strategy	Unclear 50% decrease in probability of biliary stent occlusion.
		Almadi et al (2021)(49)	No	Adult patients with borderline resectable pancreatic cancer	Up to 1 year	Metal stent: $9,304	Cost-effectiveness ratio: 9689 for metal stent and 13612 for plastic stent. Incremental cost-effectiveness ratio: −19816 for plastic stent	Choosing a metal stent at the initial insertion yields a 96 % probability for the patient to be ready for surgery. Choosing a plastic stent at the initial insertion yields an 85 % probability for the patient to be ready for surgery.
		Yoon et al (2009)(50)	No	Patients with obstructive jaundice caused by unresectable malignant biliary obstruction	End of study was either patient death or 06/2007	Metal stent: $563.98-810.86	The mean total cost for relief of jaundice was $1448.77 ± 99.52 for the metal stent group and $1319.26 ± 126.12 for the plastic stent group. Overall cost of biliary drainage was not statistically different between metal stent group and the plastic stent group.	Utilization of metal stent resulted in longer stent patency, fewer procedure-related complications, and shorter hospitalization for cholangitis.
Pillcam Colon 2 or NaviCam	Colon capsule or video capsule	Palimaka et al (2015)(51)	No	Patients aged 18 to 100-plus years who had been referred for CT colonography for known or suspected colonic disease	Life time horizon	CA $650 for the capsule Average cost of colon capsule endoscopy: CA $1,120.42	If colon capsule endoscopy were to replace CT colonography for patients aged 18 and older, the program would cost about $8.15 million in the first year, assuming complete uptake of the capsule procedure. This would amount to an additional annual expenditure of $2.72 million. If colon capsule endoscopy were to replace CT colonography for patients aged 18 and older who had undergone an incomplete colonoscopy within the previous year (about 28% of those receiving a CT colonography scan for colonic polyps), the program would cost about $6.17 million in the first year, assuming complete uptake of the capsule procedure. This would amount to an additional annual expenditure of about $741,000.	The additional cost of unnecessary colonoscopies for patients with false-positive results and additional cost and life-years lost for patients with false-negative results were used to estimate a cost-effectiveness of about $26,750 per LYG for colon capsule endoscopy versus CT colonography.
		Barré et al (2020)(52)	No	French population entering the colorectal cancer screening program: individuals aged 50 years with no prior history of screening, and no familial history of colorectal carcinoma, inflammatory bowel disease, familial adenomatous polyposis or Lynch syndrome	Unknown	Cost for capsule: 152.61 euros	Incremental cost-effectiveness ratio in the capsule endoscopy group dominated	Capsule endoscopy: - 1.2% colorectal carcinoma incidence. All comparative strategies were associated with an increase in the rate of severe adverse event related to the screening or treatment of adenomas compared with fecal immunochemical test.
		Saunders et al (2019)(53)	No	Simulated patients with Crohn disease	5-20 years	Unknown	The annual saving per small bowel and colon video capsule endoscopy per patient was $1,135. Small bowel and colon video capsule endoscopy provided a savings of $6648 over common monitoring practice over the 20-year time horizon. The total savings for the population of 12,000 patients over the 5-year period were $36,465,960.	Over the 20-year time horizon, use of common monitoring practice resulted in accumulated care costs of $320,015 and patients had an average life expectancy of 17.9 years, yielding total cost per patient life-year of $18,616. Patients achieved 8.0 quality-adjusted life years. Total care costs with the use of small bowel and colon video capsule were $313,367 and patients had an average life expectancy of 18.15 years, yielding a cost of $17,265 per patient life-year. These patients accrued 8.7 quality-adjusted life years.
		Jawaid et al (2020)(54)	No	Adult patients at least 18 years old, able to provide informed consent, hemodynamically stable, and admitted for evaluation of new onset, non-hematemesis gastrointestinal bleeding	Unknown	Unknown	Per inpatient case, there was no difference in total charges to the hospital, direct costs to the hospital, or direct cost contribution margin for both the early capsule deployment and standard of care groups.	An overall total positive test result was identified in 69.2% in the early capsule deployment group compared to 32.6% in the standard of care group. A positive test result was identified on the first diagnostic procedure 69.2% of the time in the early capsule deployment group compared to only 27.9% of the time in the standard of care group.
		Lobo et al (2020)(55)	No	Simulated patients with Crohn disease	20 years	800.75 GBP	Mean annual cost of care per patient receiving standard of care is £2191. The total mean cost per patient over the 20-year time horizon was £42 266. Cost was not substantially different than with pan-intestinal video capsule endoscopy: average £1960 per patient per year with mean total costs over 20 years of £38 433.	With colonoscopy 10.67 quality-adjusted life years were accrued over the 20 years, with pan-intestinal video capsule endoscopy accruing an additional 0.29 (total 10.96) quality-adjusted life years.
		Meltzer et al (2014)(56)	No	65-year-old patient with GI bleed	Unknown	USD $700	In low-risk group, video capsule endoscopy had a cost of $5,691 which dominated other strategies: nasogastric tube strategy with cost of $8,159 and admit-all strategy with cost of $22,766 In moderate-risk group, video capsule endoscopy continued to be the preferred strategy with cost of $9,190 but no longer dominated nasogastric tube with cost of $9,487. The admit-all strategy was dominated with cost of $22,584.	In the low-risk group, video capsule endoscopy strategy, nasogastric strategy, and admit all strategy all had effectiveness of 14.69 quality-adjusted life years. In moderate-risk group, video capsule endoscopy had 14.56 quality-adjusted life years, nasogastric tube had 14.58 quality-adjusted life years, and all admit had 14.54 quality-adjusted life years.
Eclipse	No cost analysis study
IB-Stim	Percutaneous electrical nerve field stimulation	Shah et al (2024)(57)	No	Adolescent patients with IBS who have failed at least one common empirical treatment modality for abdominal pain	1 year	Unknown	Percutaneous electrical nerve field stimulation offered the potential to save $4744 to insurers compared to usual care. Percutaneous electrical nerve field stimulation offered the potential to save $5802 by requiring lower out-of-pocket healthcare expenses and fewer missed days from work.	18 healthy days (quality-adjusted life year gain of 0.05) over a 1-year period compared with usual care without percutaneous electrical nerve field stimulation.
Echotip Insight	No cost analysis study
Mucosal Integrity Conductivity (MI) Test	No cost analysis study
GI Genius	Computer- aided detection of adenoma	Barkun et al (2023)(58)	No	Patients 50 years of age and older undergoing colonoscopy as a result of positive FIT	Life time horizon	Cost for overall case: $2,990.74	If applied to 1,000 colonoscopies yearly, the per-case cost of computer-aided detection colonoscopy and conventional colonoscopy were $2,990.74 and $3,004.59, respectively. Approximately $14 overall cost savings in the intervention group.	The quality-adjusted life year gains for intervention group and conventional colonoscopy group were 17.137 and 17.113 (P = 0.024) per thousand patients.
		Hassan et al (2023)(59)	No	Simulated cohort of patients aged 50 undergoing colonoscopies after positive FIT results	Life time horizon of 70 years	Average cost of €43,500 for each GI Genius Module	Total cost per person: €1,353.75 for standard colonoscopy and €1,353.75 for colonoscopy + GI Genius. Overall, €14.34 total savings per person.	The strategy of GI Genius-assisted colonoscopy was dominant due to an improvement in both survival (+0.024 life years) and quality of life (+0.027 quality-adjusted life years) combined with a savings of €14.34 per patient.
		Thiruvengadam et al (2023)(60)	No	Patients undergoing screening colonoscopy at age 45, and continued surveillance or screening colonoscopies	Up until age 100 or death	Monthly unit cost of $3000 with 40 colonoscopies per unit for a per-procedure cost of $75	Scenario 1 (real-time computer-aided detection improves the adenoma detection rate of all endoscopists): using computer-aided detection had overall cost saving of $143 per person. Scenario 2 (real-time computer-aided detectionsets a minimum adenoma detection rate for colonoscopy performance): using computer-aided detectionhad overall cost saving of $7 per person. Scenario 3 (real-time computer-aided detection improves the adenoma detection rate of all physicians but only up to a defined maximal level/ceiling): overall cost saving of $108 per person.	Scenario 1 (real-time computer-aided detection improves the adenoma detection rate of all endoscopists): 30.4% fewer lifetime colorectal cancer cases, 32.7% fewer colorectal cancer deaths, and an additional 0.01 quality-adjusted life years. Scenario 2 (real-time computer-aided detection sets a minimum adenoma detection rate for colonoscopy performance): 21.5% reduction in colorectal cancer incidence, a reduction of 24.0% in colorectal cancer mortality Scenario 3 (real-time computer-aided detection improves the adenoma detection rate of all physicians but only up to a defined maximal level/ceiling): reduction of an 28.5% in colorectal cancer incidence, a reduction of 29.8% in colorectal cancer mortality.
AccuMeasure	No cost analysis study
Vibrant	No cost analysis study
Acticon Neosphincter	No cost analysis study
Axonics or Interstim	Sacral neuromodulation or sacral nerve stimulator	de Miguel Valencia et al (2023)(61)	No	Patients with severe fecal incontinence between who did not respond to previous treatments	Unknown	Cost/unit: 3354.59€	The mean cost of treatment with sacral neuromodulation was 3544€/patient/year, vs the estimated mean cost of the standard of treatment was 1097€/patient/year.	After definitive implantation, there was a significant decrease in the mean number of days per week with an incontinent episode, from 4.98 to 1.25, and in mean Wexner's score from 16.88 to 6.95. Treatment with sacral neuromodulation showed significant improvement in quality of life.
		van Wunnik et al (2012)(62)	No	Theoretical patients with fecal incontinence	5 years	Cost of device: €12698	The cost per successfully treated patient in the sacral neuromodulation− strategy was €38392 and €20113 for sacral neuromodulation+ strategy.	The sacral neuromodulation− strategy resulted in a success rate of 0.59, whereas for the sacral neuromodulation+ strategy was 0.82 after 5 years. Sacral neuromodulation− yielded 4.14 quality-adjusted life years and sacral neuromodulation+ had a total number of 4.21 quality-adjusted life years.
		Hounsome et al (2018)(63)	No	Patients with fecal incontinence who failed conservative therapy	5 years	Total temporary sacral nerve stimulator cost: £1,612.94 Total permanent sacral nerve stimulation cost: £12,216.43	The estimated savings from offering percutaneous tibial nerve stimulation first were £662–£5,697 per patient.	Intervention+sacral nerve stimulation strategy was less costly and more effective compared with sacral nerve stimulation alone. The probability of intervention+sacral nerve stimulation being cost-effective was around 80% within the cutoff range of £30,000 per quality-adjusted life year.
		Tipsmark et al (2016)(64)	No	Patients with diarrhea-predominant or mixed irritable bowel syndrome (IBS) according to the Rome III criteria and a baseline Gastrointestinal Symptom Rating Scale – IBS version score of at least 40 points	4-20 years	£9,919	The extra cost of sacral nerve stimulation for irritable bowel syndrome vs no treatment was £5,085 per patient over a 4-year perspective and £2,476 per patient over a 20-year perspective Incremental cost-effectiveness ratio: £31,270/quality-adjusted life year after 4 years and £18,901/quality-adjusted life year after 20 years. Cost-effectiveness was not reached for 4 years after permanent implantation but may become cost effective over a 7-year perspective.	The corresponding utility gain was 0.163 quality-adjusted life year and 0.131 quality-adjusted life year per patient.
		Indinnimeo et al (2010)(65)	No	Patients with fecal incontinence	5 years	€2,500 for sacral neuromodulation test €10,000 for final implantation	For patients with a structural deficiency, sacral neuromodulation generated an additional cost of €2,417 over the 5-year period, resulting in an incremental cost-effectiveness ratio of €5,374 per symptom-free life-year gained. For patients with intact sphincters, sacral neuromodulationgenerated an additional cost of €4,878, resulting in an incremental cost-effectiveness ratio of €7,346 per symptom-free life-year gained.	The corresponding quality-adjusted life year gains were 0.085 for patients with a structural deficiency and 0.13 for patients with intact sphincters.
		van der Wilt et al (2017)(66)	No	Children and adolescents aged 10-18 years with constipation refractory to conservative management	3 years	€8,739 for device and €2,303 for tined lead	The mean cumulative costs per patient over the 3-year period in the conservative treatment and sacral neuromodulationgroups were €7,574 and €17,789, respectively.	Mean effectiveness per patient over the 3-year period in the conservative treatment and sacral neuromodulation group were 0.86 quality-adjusted life years and 1.74 quality-adjusted life years , respectively.
		Hetzer et al (2006)(67)	No	Patients who had more than one episode of fecal incontinence per week for at least a year and failed medical therapy	Unknown	€15,345 per patient for the first year and €997 annually thereafter	Compared with other surgical treatment options, sacral nerve stimulation is more cost-effective in the short term, although long-term results are still pending. However, conservative treatment and anterior anal sphincter repair were noted to be less expensive.	With conservative treatment, quality of life is significantly reduced and a symptomatic recurrence rate of up to 50% in the long term is described after anal sphincter repair. The comparison with the anal sphincter repair group is limited in that some of the patients in the sacral nerve stimulation group would not have been suitable for an anal sphincter repair. Sacral nerve stimulation can be performed with low operation and patient care costs, but complications are less frequent and less severe, and therefore less expensive.
		Dudding et al (2008)(68)	No	Adult patients with fecal incontinence who have failed conservative treatment	Unknown	Cost per device: £7,896·00	For the basic model, which excluded indirect non-medical costs, the decision tree favored the conservative strategy (cost-effectiveness ratio of £596 per quality-adjusted life year as opposed to £2,159 per quality-adjusted life year for the sacral nerve stimulation strategy). The incremental cost-effectivenss ratio was £25,070 per quality-adjusted life year gained. When a patient-year of full continence was used as the unit of effectiveness, the cost-effectiveness ratio for sacral nerve stimulation was £4,487 per patient-year of continence and the incremental cost-effectiveness ratio was £3,329 per patient-year of continence gained. When patients with greater than 50% improvement in continence were also included in the analysis, the cost-effectiveness ratio was calculated to be £2,079 with an incremental cost-effectiveness ratio of £1,543 per patient-year of improved continence gained.	All 70 patients underwent temporary sacral nerve stimulation, 61 (or 8%) had over 50% reduction in incontinence episodes and were therefore deemed eligible for permanent implantation. At latest follow-up, a median of 24 months after permanent neurostimulator implantation, continence had improved in 41 of the 48 patients. The total number of incontinence episodes was reduced from a median of 6 (range 0–81) to 0·5 (range 0–59) episodes per week.
Linx	Magnetic sphincter augmentation	Harper et al (2023)(69)	No	Patients who were refractory to PPI therapy	Life time horizon	Cost per device: £3,000	Unknown	RefluxStop was more cost effective and had favorable surgical outcome compared to magnetic sphincter augmentation
		Pandolfino et al (2020)(70)	No	Simulated patients with gastroesophageal reflux disease refractory to medication management	1 year	Inpatient: $21,817 Outpatient: $8,273	The introduction of magnetic sphincter augmentation would lead to a cost savings of $111,367 annually for a commercial plan. 0.01 per member per month savings for 1 year.	Superior side-effect profile, improved moderate-to-severe regurgitation results, and similar or better health-related quality of life with magnetic sphincter augmentation versus laparoscopic Nissen fundoplication for treating medication-refractory patients.
OrganCare System or OrganOx	Normothermic machine perfusion	Wehrle et al (2024)(71)	No	Patients undergoing primary deceased donor liver transplantation	Unknown	Organ acquisition and preservation with normothermic machine perfusion: $88,923	Organ acquisition/preservation was more expensive with normothermic machine perfusion ($88,923 vs $52,446), yet cumulative costs were not different ($221,952 vs $219,049).	Normothermic machine perfusion-treated donation after circulatory death-grafts experienced a reduction in the median Comprehensive Complications Index in the 90-day post-liver transplant period compared with the static cold storage cohort as well as a reduced intensive care length-of-stay. There was no difference in Comprehensive Complications Index for donation after brain death-grafts based on normothermic machine perfusion for the inpatient or 90-day periods. Normothermic machine perfusion treatment was associated with a reduced rate of post liver transplant renal replacement therapy, relaparotomy, and bleeding requiring reintervention.
		Webb et al (2022)(72)	No	Patients undergoing liver transplantation	5 years	Normothermic machine perfusion per run is $15,359 to $16,720	The mean cost for the normothermic machine perfusion strategy was $456,455 versus $519,222 for the control strategy.	The incremental effectiveness of normothermic machine perfusion was 3.48 quality-adjusted life years versus 3.17 for the control. The higher number of quality-adjusted life years accumulated in the normothermic machine perfusion strategy is from the additional livers transplanted from the OrganOx machine, which leads to a subsequent decrease in both waitlist death and patients awaiting transplantation.
		Javanbakht et al (2020)(73)	No	Hypothetical cohort of patients awaiting liver transplantation	Life time horizon	Cost of disposables and solutions: £6,000	The total costs per patient were £37,370 vs £46,711 for SCS and normothermic machine perfusion groups, respectively. The estimated incremental cost-effectiveness ratio was £7,876 per each quality-adjusted life year gained.	Effectiveness per patient was 9.09 QALYs vs 10.27 QALYs for SCS and NMP groups, respectively.
Solesta	Fecal incontinence bulking agent	Bernstein et al (2014)(74)	No	Adult patients who have had fecal incontinence for at least 12 months and did not respond adequately to conservative therapy	Unknown	Cost of device: $4,900	The incremental costs per quality-adjusted life year gained were $37,036 for non–animal stabilized hyaluronic acid and dextranomer copolymer versus conservative therapy and $244,509 for sacral nerve stimulation versus non–animal stabilized hyaluronic acid and dextranomer copolymer. The incremental cost-effectiveness ratios per additional incontinence free days gained were $42.60 for non–animal stabilized hyaluronic acid and dextranomer copolymer versus conservative therapy, $73.76 for sacral nerve stimulation versus non–animal stabilized hyaluronic acid and dextranomer copolymer, and $62.55 for sacral nerve stimulation versus conservative therapy.	Total incontinence free days during the 3-year periods associated with conservative therapy, non–animal stabilized hyaluronic acid and dextranomer copolymer, and sacral nerve stimulation were 128.8, 267.6, and 514.8 days, respectively.
Elra Electrode	No cost analysis study

Cost-effectiveness

Among the 19 devices with clinical trial data, twelve devices (63.2%) had published cost-effectiveness data. A total of 35 cost-effectiveness analyses were identified (Table 3 for published cost-effectiveness data available for each gastroenterology device). 27 (79.4% of 35) analyses demonstrated the device to be more effective than the alternative device or treatment strategy. Out of these 27 studies, 12 showed that the devices were more costly, 12 showed that the devices were less costly, and 3 showed that the devices were of similar cost to their compared devices or treatment strategy. Five studies (14.7%) showed that the device was less effective when compared to their alternative devices, with four devices being more costly and one device was of unknown cost when compared to their alternative device or treatment strategy. Lastly, two studies (5.8%) showed that the studied device demonstrated similar effectiveness when compared to the alternative device or treatment, but was notably at lower cost.

Cost-effectiveness analyses of new medical devices can help healthcare systems navigate how to maximize health outcomes relative to available resources.(31) Of the devices with clinical trial data, the majority demonstrated increased effectiveness when compared to the alternative device or treatment strategies. While nearly half of these devices were more costly upfront, their increased effectiveness may eventually offset these initial costs by reducing indirect medical costs, such as those incurred from treating residual symptoms, time lost from work, or additional required procedures. However, there were a few studies which did show that these novel devices were not only more costly, but were also less efficacious. Practicing gastroenterologists should be aware of the studied indications in these cost-effectiveness analyses and whether these studies can be applied to their patient populations. In addition, such studies can further guide risk-benefit discussions with patients when considering utilization of such novel medical devices.

Barriers and realities in innovation

The worldwide regulatory and intellectual property protection climate constantly evolves, and with that, where the newest innovations might originate across the world. For example, the FDA process enables marketing in the United States while CE mark enables marketing in the European Union (EU).(32) The barrier to EU entry comparatively increased with new Medical Device Regulation (MDR) requirements for clinical data in 2017 [(EU) 2017/745]. Whether the intended outcomes of these increased requirements were successfully achieved has remained unclear.(33) The costs and perceived rigor of intellectual property protection across the world also affects the likelihood and timing in which industry will bear the risk of bringing new technologies to market.

With these considerations in mind and in contrast to years past, the United States may be positioned as the first market to receive new medical technologies over the next decade. Recent examples of devices that come to market without needing clinical data include the Spyglass DS and DS II direct visualization system to diagnose and treat biliary duct disease and the Overstitch device that has enabled endoscopic sleeve gastroplasty. Neither of these devices appear in our tables, because FDA did not require clinical data in their 510(k) applications, yet both of these devices have successfully reshaped the field of advanced endoscopy.

Trends and directions in the past decade

In the last decade, almost two-thirds of new medical technologies (n = 442 or 62.0%) in gastroenterology have been developed for use in general endoscopy. Most gastroenterology technologies came to market through the 510(k) regulatory pathway for devices, in which the FDA clears a device as being substantially equivalent to an existing device and with low or moderate risk to patients. Over 98% of these 510(k) devices came to market without clinical data. Among new technologies that were first-of-its-kind in terms of intended use or technology, clinical trials were generally performed to support FDA approval and published in the medical literature. Notably, of those devices with clinical trials, only a small minority were randomized control trials with adequate sample size. Adverse events were also common, especially in devices with first-in-class and first of intended use technologies.

Gaps in knowledge

The quality of clinical evidence inherently differs between medical technology trials and drug trials. For example, new drugs are expected to be supported by two phase III randomized clinical trials powered to detect change in a primary outcome measure. Patient-reported outcome measures should also be incorporated into primary outcome measures in drug trials. Medical technology trials that are successfully incorporated into regulatory applications do not necessarily require binary endpoints as might be expected in drug trials.(34) For medical technology, over 85% of trials did not use a randomized, placebo-controlled design and were not adequately powered to detect changes in clinical outcomes. Only two trials were powered by RCT and included patient-reported outcome assessments in the primary endpoint.(16,20) Our findings contrast with physician expectations; in one study, 35% of physicians believed that medical technology approvals were based on two or more randomized controlled trials, 66% thought they were based on a single RCT, and 12% felt that they were approved based on registries or other nonrandomized studies.(1) Medical technology trials might also use methods other use binary endpoints as might be expected in drug trials.(34)

This discrepancy broadly relates to differences in the regulatory bar for new medical technology compared to new drugs. For first-of-its-kind medical technology or high-risk technology, the FDA weighs probable benefit against any probable injury or illness related to the technology, in contrast to more stringent definitions of safety and effectiveness in drug regulation that demand more extensive clinical trials. Where this becomes crucial is in understanding the nature of pathology that new technology addresses, with a prime example being the poo stent to treat pancreatic walled-off-necrosis. Regulators consider the inherent complication rates and natural history if left untreated. along with comparison to other available alternatives such as open or laparoscopic surgical necrosectomy. The adverse events from using the Axios stent might seem high at first glance without proper context to the risk reduction offered by this device compared to alternatives.(35) Unlike drugs, most lower-risk technologies are cleared as substantially equivalent without needing new clinical data instead of being approved based on safety and effectiveness.(36) As with drug trials, clinical data for medical technology (when available) are relevant only to the target population and according to intended use and adequate instruction.

The FDA can require post-marketing studies or clinical registries to address this knowledge gap and capture information on safety and effectiveness. Robust post-marketing surveillance mechanisms are used at the FDA to monitor device safety, but their effectiveness hinges on gastroenterologists reporting adverse clinical or technical performance events.(37) Lastly, the FDA does not regulate the practice of medicine. The FDA’s stance on off-label use of devices is that “If physicians use a product for an indication, not in the approved labeling, they have the responsibility to be well informed about the use and effects of the product.”(38)

Future directions in the next decade

The shortage of clinical data for new medical technologies does not imply that such data are unnecessary. Recognizing that large placebo-controlled medical technology trials might be unviable, the FDA allows qualitative or quantitative patient preference studies to inform the comparative assessment between probable benefit and risk.(39) Such studies can be beneficial in cases of uncertainty and medical device industry should consider the needs of gastroenterologists and patients in designing outcome measures relevant to practice. Finally, gastroenterologists and patients should be aware that majority of clinical devices in use were approved or cleared without clinical data, and its use largely hinge on the experience and familiarity of the practicing gastroenterologist. Therefore, future trials for new medical technology should consider the probable benefits and risks that are of interest to practicing gastroenterologists and patients, rather than focusing on the core requirements of regulators.

Framework to consider medical technology in development and clinical practice

From a technology development standpoint, a unified framework with the needs of clinical practice ensures that new medical technologies are designed from the start to reflect how they will eventually be used (Figure 1). From a gastroenterologist’s standpoint, a shared framework enables their practice to appropriately mitigate the risk of being an early adopter of new technology. When safety data are limited, clinicians should transparently consider, communicate and document the risks, benefits and alternatives with patients. Given alternatives that can be unfavorable (such as in our example of open necrosectomy), patients might reasonably choose care options informed directly by the clinician’s own experience. FDA’s Manufacturer and User Facility Device Experience (MAUDE) database is the most up-to-date resource for gastroenterologists to look up and report device complications, and it inherently relies on gastroenterologists to report adverse events to industry or directly to FDA.(40) Newer procedures may not be covered by insurance, therefore practices should ask industry representatives to present the business case adapted to the local practice to justify cost equity. Practices should be equipped to routinely handle pre-authorizations for unlisted procedures (e.g. Computerized Procedural Terminology code 91299) along with supporting clinical evidence.(41) To assist others in supporting insurance coverage, gastroenterologists should monitor and publish on their own real-world outcomes and safety events regardless of perceived positive or negative results.

Figure 1:

Framework to evaluate new medical technologies from patient, gastroenterology and society perspectives.

TAKE-HOME POINTS

In conclusion, the gastroenterology pipeline is robust and expanding within and beyond the endoscopy suite. The lack of clinical data before the market availability of new medical technologies should be considered in context: regulators deem most devices to be no more than minimal risk, and device/technology trials are inherently different from drug trials. This reality stresses the balance between reasonable timing to have access to beneficial technologies on the market without delaying care with a reasonable clearance process to ensure safety given that the clinical real world experience have many confounders and variables unlike regulated conditions in laboratories and clinical trials. Our findings should also encourage gastroenterology innovators to think “outside the box” in demonstrating that successful gastroenterology innovations can absolutely have various levels of clinical use and reach in practice.