Contents
On this page
- Differences between device evidence and drug evidence
- Device iteration and identification of device generation
- The operator as a component of the intervention
- Registries, regulatory files and surveillance data
- Heterogeneity of outcome definitions across studies
- Items commonly omitted from device reviews
- Appraisal of predominantly non-randomised device evidence
- Pooling, survivorship and competing risks
- Applicable reporting standards
- Key points
- Application to your study
- Frequently asked questions
Academy · Research Data Analytics & AI · implantable device reviews
Reporting Quality in Systematic Reviews of Implantable Devices
Systematic reviews of implantable devices are often written as though the device were a drug, and the items they consequently fail to report are those that determine whether the result is transferable: the model and generation implanted, the operator and their experience, the surgical technique, the definition of failure and the duration of follow-up. A review that reports a pooled revision rate without these items produces a figure with no clear referent.
Device evidence is structurally different from drug evidence. The implant changes between generations while the literature accumulates, the operator is part of the intervention, the comparator is often a previous device rather than a placebo, randomised trials are the exception, and the most informative data are held in registries and regulatory files rather than journals. These features do not make a review impossible, but they mean that the reporting items that matter differ from those a drug-review template requires, and the gap between the two lists is where device reviews lose their usefulness.
Learning objectives
- Description of the intervention in a device review at the level of detail that determines transferability
- Accounting for device iteration, so that the review states which generation its result applies to
- Treatment of operator experience and procedural volume as effect modifiers rather than noise
- Use of registries, regulatory documents and post-market surveillance as named sources
- Harmonisation of outcome definitions (revision, failure, explant and survivorship), or reporting that harmonisation was not possible
- Assessment of whether pooling is defensible, and correct presentation of survivorship evidence when it is not
Common misconceptions
- “The device is the intervention.” The intervention is a device model and generation, implanted by a particular operator using a particular technique, with a particular rehabilitation or follow-up protocol. Reporting only the device name describes a fraction of the intervention delivered.
- “Absence of randomised trials means the evidence is weak.” It indicates that randomised evidence is sparse, which is a different statement. Large registry series contain information about long-term failure that no trial of feasible duration would capture, and a review that dismisses them reports less than it could.
- “Revision rate is a standard outcome.” Revision, reoperation, failure, explant and loss of survivorship are defined differently across studies, and pooling them as though they were one outcome produces an uninterpretable figure.
- “A pooled estimate is the goal.” For time-to-event device outcomes with differing follow-up, a pooled proportion is usually inappropriate. Survivorship at stated timepoints, with numbers at risk, is more informative and makes more modest claims.
- “PRISMA covers what a device review needs to report.” PRISMA governs the reporting of the review. Device model and generation, operator experience, technique, implant positioning and failure definitions are study-level items, and they must be extracted and reported deliberately.
Differences between device evidence and drug evidence
Five differences alter what a review must do, and all five are reporting problems before they are analytical ones.
The intervention is a composite. A device, a surgical approach, an operator, an implant position, a perioperative protocol and a rehabilitation pathway jointly produce the outcome. Any of these can differ between two studies that name the same device.
The device changes. Devices iterate (materials, geometry, coatings, fixation, sizing), and the literature on a product line accumulates across generations that are not equivalent.
Blinding is usually impossible and randomisation often is too. Comparative evidence is frequently observational, and the comparison is often with a previous device rather than with no device 28.
Outcomes are long-horizon and competing. Failure may occur years after implantation, and patients die of other causes in the meantime, which makes naive proportions misleading and competing-risk treatment relevant.
The most complete data are often not in journals. National and regional registries, regulatory submissions and post-market surveillance reporting hold long-term follow-up data at a scale that the published literature does not 14,25. Our pages on regulatory challenges for devices and device and IVD evidence support address the regulatory context from which these documents originate.
Device iteration and identification of device generation
This is the question that most device reviews cannot answer about themselves, and it is the first that a reader facing a decision will ask.
A product line implanted over fifteen years may have changed its bearing material, fixation method, coating, instrumentation and sizing range. Studies published across that period report on different objects under one trade name. Pooling them yields an average over a product history, which is informative about the past and potentially misleading about the implant currently available; moreover, the direction of the error is not predictable, since a later generation may perform better or worse.
Three steps are required. First, the model, generation, catalogue or reference identifiers (where reported) and the implantation date range should be extracted for every included study; the implantation period is often the only available proxy for generation when model detail is absent. Second, the number of included studies that permitted a generation to be identified at all should be reported, since that count is itself a finding about the literature. Third, the conclusion should state which generation the result applies to, or state that this could not be determined. “These findings relate to implants placed between the stated years and may not apply to the current generation of this device” is a sentence that most device reviews should contain and few do.
Where a device has been subject to a field safety notice, a recall or a design change following adverse outcomes, that history should be reported in the review, because it bears directly on which studies describe the implant a reader is considering.
The intervention description in a device review
Describes a trade name, not an intervention
We included studies of patients who underwent implantation of the device for the stated indication, and extracted the revision rate at final follow-up.
“The device” names a product line, not an implant. A trade name may cover generations with different materials, geometry and fixation, which do not constitute the same intervention.
No implantation dates. Without model detail or a date range, it is impossible to determine which version of the product the result describes.
No surgical or technical detail. Approach, fixation and positioning are part of the intervention and can influence the outcome independently of the device.
Operator and centre invisible. Learning-curve and specialist-centre results will be pooled with general practice as though they estimated the same quantity.
“Revision rate” with no definition. Revision, reoperation, failure and explant are defined differently across studies, so the extracted quantity is not a single quantity.
“At final follow-up” is not a timepoint. Two years and ten years are combined, without numbers at risk, so the proportion has no time referent.
No manufacturer involvement recorded. In a literature with substantial industry support, the distribution of such involvement across the included studies is information a reader requires.
Describes what was done, to whom and by whom
For each included study we recorded the manufacturer, the device model and, where reported, the generation and catalogue or reference identifiers, together with the range of implantation dates, which we used as a proxy for generation where model detail was absent. We recorded the surgical approach, the fixation or attachment method, implant sizing or configuration where reported, and the perioperative and rehabilitation protocol. We recorded operator experience with this specific device, centre procedural volume, whether the study period included the device’s introduction at that centre, whether any author was a developer of the device or practised in a designated centre of excellence, and whether a proctoring or training programme was described. For outcomes we recorded each study’s own definition verbatim — distinguishing revision, reoperation, clinical failure, radiographic failure and explant — together with minimum, median and maximum follow-up and the number remaining at risk at each reported timepoint, and we report the mapping we applied to group these definitions for synthesis as a published table. We recorded funding source and any manufacturer involvement in design, conduct, analysis or authorship. Where an item was not reported we recorded it as not reported rather than inferring it, and we report the proportion of included studies supplying each item. Our conclusions are stated as applying to the generations and care contexts represented by the included studies, which are specified, and we state explicitly where the generation could not be determined.
The implant is identified at the level of model and generation where possible, with implantation dates as the fallback, so that the result can be attributed to a specific product.
Technique, fixation and positioning are captured, which keeps the composite nature of a device intervention visible rather than reducing it to a name.
Operator experience, centre volume and developer involvement are extracted, making them available as pre-specified subgroups rather than post hoc explanations.
Outcome definitions are recorded verbatim and the grouping decision is published as a mapping table, so that a reader can see exactly what was treated as equivalent.
Follow-up is recorded with numbers at risk, which allows survivorship to be reported accurately rather than as a single undated proportion.
Unreported items are recorded as unreported and counted, so that the extraction yields a transferable statement about reporting practice in this device area.
The conclusion is explicitly restricted to the generations and settings covered by the evidence, which is what makes a device review usable in a decision.
A clinician reading the first version obtains a figure; a clinician reading the second learns whether that figure applies to the implant about to be used, placed in the manner the clinician places it, by an operator of comparable experience. This is the essential difference between a device review that informs a decision and one that cannot.
The operator as a component of the intervention
For implantable devices, the outcome depends on the implanting clinician and on the surrounding system of care. Learning curves are well recognised: early cases in an operator’s or a centre’s experience with a device frequently differ from later ones, and a study conducted during adoption measures something different from a study in established practice.
This has consequences on which a review can act. The following should be extracted where reported: operator experience with the specific device, centre procedural volume, whether the study period covers the introduction of the device at that centre, whether the operators were the developers of the device or practised in a designated centre of excellence, and whether a proctoring or training programme was in place. Results from developer series and high-volume specialist centres are not estimates of outcomes in general practice, and a review that treats them as such overstates performance.
Where these items are unreported, which will usually be the case, this should be recorded. A review reporting that a minority of included studies described operator experience has made a useful statement about the evidence base and has supplied the reason why its own pooled estimate should be interpreted cautiously. Where sufficient studies report them, operator experience and centre volume are candidates for pre-specified subgroup analysis rather than post hoc explanation, and should therefore be declared in the protocol. Our page on framing the review question describes how to specify an intervention that includes the person delivering it.
Registries, regulatory files and surveillance data
A search restricted to journals for a device question omits the sources with the longest follow-up and the largest numbers 14,18. Four additional source types should be specified as search steps.
Device registries. National and regional arthroplasty, cardiac device, vascular and other registries publish annual reports with survivorship by implant, sometimes by brand and generation. They are not peer-reviewed journal articles, yet they frequently constitute the best available long-term evidence. The protocol should state whether registry reports are eligible as included evidence, as contextual evidence, or not at all 5.
Regulatory submission and approval documents. Pre-market submission summaries, approval documents and public summaries of safety and clinical performance contain study data, and for devices they often contain the only publicly available report of a pivotal study 25. See technical file and submission support and summaries of safety and clinical performance.
Post-market surveillance and follow-up reports. Periodic safety reporting and post-market clinical follow-up generate structured outcome data over the time horizon in which implant failures occur. See post-market surveillance and post-market clinical follow-up.
Adverse event databases and field safety notices. These sources provide no denominators and cannot support rate estimates, which is precisely why they should be used for signal detection and history rather than for incidence; a review should state this distinction explicitly when it uses them.
The general structure of a multi-tier search is described on our page on sources of review data.
Heterogeneity of outcome definitions across studies
Device studies report outcomes that appear equivalent but are not. The most common conflations are set out below.
Revision usually denotes a further procedure involving the implant, but studies differ on whether a procedure that exchanges only a component counts, and on whether a reoperation that does not involve the implant counts. Reoperation is broader and includes procedures unrelated to implant performance. Failure may be defined clinically, radiographically, functionally or by the occurrence of revision, and a radiographic failure is not a revision. Explant is narrower than revision. Survivorship is a time-to-event quantity whose value depends on the endpoint used to define it, the follow-up achieved, and the handling of censoring and competing mortality.
Three practical rules apply. First, the definition should be extracted verbatim for every study rather than mapped to the reviewer’s own category at extraction, because mapping is an analytical decision and should be visible as such 2,6,17. Second, studies should be grouped for synthesis only where the definitions are genuinely compatible, and the mapping table should be reported so that a reader can see what was treated as equivalent. Third, where the definitions cannot be reconciled, this should be reported as the result for that outcome rather than pooling across them; an accurate statement that the literature does not define failure consistently is more useful than an average of incompatible quantities.
Follow-up duration requires the same treatment. Minimum, mean or median and maximum follow-up should be extracted with the number still at risk, because a revision proportion at two years and one at ten years are not the same outcome. Our page on what data to extract covers the general practice of recording the five elements of any outcome 2,26.
Outcomes that appear equivalent but are not
| Outcome term | What it usually means | Where studies differ | Consequence of pooling across the difference |
|---|---|---|---|
| Revision | A further procedure involving the implant | Whether exchanging a single component counts; whether a reoperation not touching the implant counts | A revision rate that includes and excludes component exchanges in the same estimate |
| Reoperation | Any further procedure at the site | Includes procedures unrelated to implant performance | Device performance inflated by unrelated surgery |
| Failure | The implant has not achieved its purpose | Defined clinically, radiographically, functionally, or by revision having occurred | Radiographic findings and surgical events counted as one quantity |
| Explant | Removal of the device | Narrower than revision; sometimes reported as a subset of it | Underestimation if read as revision, overestimation if the reverse |
| Survivorship | Time-to-event freedom from a defined endpoint | The endpoint used; follow-up achieved; censoring and competing mortality handling | Survivorship figures from different endpoints and horizons presented as comparable |
Compiled from the definitional variation found across device outcome reporting; each study’s definition should be extracted verbatim and the mapping applied should be published.
Items commonly omitted from device reviews
The reporting quality of a body of device reviews, or of a single review, is best assessed against a fixed item list. The items below are those whose absence most limits what a reader can do with the review, and they are omitted sufficiently often that counting them is worthwhile in its own right.
At study level: device manufacturer, model and generation; catalogue or reference identifiers; implantation date range; surgical approach and technique; implant position or sizing; operator experience and centre volume; perioperative and rehabilitation protocol; the verbatim definition of each outcome; follow-up duration with numbers at risk; and funding source and the role of the manufacturer.
At review level: whether registry and regulatory sources were searched; how device generations were handled in the synthesis; whether operator experience was pre-specified as a subgroup; whether competing risks were addressed; and a statement of which generation and which practice setting the conclusions apply to.
Two items merit particular emphasis because they are both consequential and inexpensive to report. The first is manufacturer involvement (funding, authorship, provision of devices or statistical support), because the device literature contains a high proportion of industry-supported studies and a reader needs to see its distribution across the included set. The second is the proportion of included studies that reported each of the items above, which converts the extraction work into a transferable statement about reporting practice in that device area.
Device-specific reporting items, and what their absence costs the reader
| Item | Level | What a reader cannot do without it |
|---|---|---|
| Manufacturer, model and generation | Study | Tell whether the result applies to the implant currently available |
| Catalogue or reference identifiers | Study | Match the studied implant to a specific product configuration |
| Implantation date range | Study | Infer generation when model detail is missing, which is the usual case |
| Surgical approach and technique | Study | Judge whether the procedure resembles their own practice |
| Implant position, sizing or configuration | Study | Assess whether a technical factor rather than the device drove the outcome |
| Operator experience and centre volume | Study | Distinguish device performance from adoption-phase or specialist-centre performance |
| Perioperative and rehabilitation protocol | Study | Separate the device’s contribution from the pathway around it |
| Verbatim outcome definitions | Study | Know whether two studies’ revision rates mean the same thing |
| Follow-up duration and numbers at risk | Study | Compare proportions that were measured over different horizons |
| Funding and manufacturer involvement | Study | See the distribution of industry support across the evidence base |
| Registry, regulatory and surveillance sources searched | Review | Judge whether the longest follow-up evidence was looked for |
| Handling of device generations in the synthesis | Review | Interpret a pooled estimate that may average across a product history |
| Competing risks addressed | Review | Trust a failure proportion in a population with substantial other-cause mortality |
| Statement of which generation and setting the conclusion applies to | Review | Apply the review to a decision at all |
Compiled as a working reporting item list for reviews of implantable devices; the proportion of included studies supplying each item should be reported, as it is itself a finding about the literature.
Appraisal of predominantly non-randomised device evidence
Most device reviews include non-randomised comparative studies and single-arm series, and the appraisal must be designed for such evidence rather than apologise for it 20,28.
An instrument designed for non-randomised studies of interventions should be used and named with its version 3. The domains most relevant to device evidence are confounding, particularly by patient selection, indication and operator; selection of participants into the study, including whether the series was consecutive; classification of the intervention, which for devices means whether the model and generation were documented; deviations from the intended intervention, including intraoperative conversion to another device; missing data, which for long-horizon implant outcomes means loss to follow-up and whether those lost differ systematically; and measurement of the outcome, which for radiographic or clinician-assessed failure is assessor-dependent.
Two device-specific considerations should be added explicitly. Single-arm series without a comparator can describe experience but cannot support a comparative claim, and a review should not allow a narrative comparison between separately reported series to become an implicit effect estimate. In addition, developer and designated-centre series should be identified as such in the appraisal, because their results are not estimates of general practice even when the studies are internally sound.
Where randomised evidence exists, it should be appraised with a randomised-trial instrument and reported separately rather than the appraisals being merged. Combining designs in a single appraisal column obscures the structure of the evidence base, which is one of the more useful outputs of the review.
Pooling, survivorship and competing risks
Four conditions should be checked before any meta-analysis of device outcomes: the outcome definitions are compatible across studies; the follow-up durations are comparable, or the analysis accounts for their difference; the device generations are the same, or generation is handled explicitly as a grouping or subgroup variable; and the care context (operator experience, centre type) is sufficiently similar that the pooled estimate refers to an identifiable practice setting.
If these conditions hold, standard methods apply; see meta-analysis support and meta-analysis methodology. If they do not, the alternatives are preferable to a pooled estimate with caveats. Survivorship should be reported at stated timepoints with numbers at risk, which respects the time-to-event structure and makes differing follow-up visible. A per-study table of revision or failure proportions should be presented with the definition, follow-up and generation in the same row, so that incomparable figures cannot be compared inadvertently. The synthesis should be grouped by generation or by care context where the data allow. Competing mortality should be identified as a methodological issue, because in an elderly implant population a patient who dies has not had a successful implant in any sense that a simple proportion captures.
Where cost or value questions follow from the clinical synthesis, they require a dedicated framework rather than an extension of the pooled estimate; see cost-effectiveness analyses for device manufacturers.
Applicable reporting standards
The review should be reported in accordance with the PRISMA 2020 statement, and the designs of the included studies should be reported against the appropriate standards: a randomised-trial reporting checklist for trials, an observational reporting checklist for cohort and case-control designs, and a structured intervention description framework for the intervention itself 11. The last of these is more important for devices than for almost any other type of intervention, because the intervention is a composite and a structured description is what makes it comparable across studies; see the intervention description checklist, the observational study checklist and the randomised trial checklist 24.
The device-specific review-level statements should then be added: which generations the evidence covers and which the conclusion applies to; whether registries, regulatory documents and surveillance sources were searched, and what they contributed; how outcome definitions were mapped, with the mapping table published; the distribution of manufacturer involvement across included studies; and the proportion of included studies reporting each key device item, which is the most reusable output of the review and the element a subsequent researcher is most likely to cite.
A final statement, usually missing, should set out what a reader cannot conclude. Most device reviews cannot tell a clinician how the current generation of a device will perform in that clinician’s hands and centre, and stating this is preferable to allowing a pooled revision rate to be read as though it could. Reporting checklist review catches the omissions before submission.
Key points
Key points
The intervention should be treated as a composite: a device model and generation, a surgical technique, an operator with a given level of experience, and a care pathway surrounding the procedure. Manufacturer, model, generation and implantation date range should be extracted for every study, together with the number of studies that permitted a generation to be identified at all, and the review should state which generation its conclusion applies to, or that this could not be determined. Each study’s outcome definition should be extracted verbatim, because revision, reoperation, failure, explant and survivorship are not interchangeable, and the mapping used to group them should be published. Operator experience, centre volume and developer involvement should be recorded, since adoption-phase and specialist-centre results are not estimates of general practice. Registries, regulatory submissions and post-market surveillance should be searched as named steps, and adverse event databases used for signal detection rather than incidence estimation because they have no denominator. Before pooling, the compatibility of definitions, follow-up, generations and care contexts should be checked; where they are incompatible, survivorship should be reported at stated timepoints with numbers at risk, competing mortality should be addressed, and the review should state what it cannot tell a clinician.
Application to your study
Application to your study
- The intervention should be recorded as five fields rather than one. Device and generation, technique, operator experience, implant configuration and surrounding care pathway.
- Implantation date ranges should always be extracted. They are the only available proxy for generation when model detail is missing, which is the usual case.
- Outcome definitions should be extracted verbatim. The mapping table showing how definitions were grouped should then be published, and definitions that cannot be reconciled should not be pooled.
- Follow-up should be recorded with numbers at risk. Minimum, median and maximum, per timepoint; a proportion “at final follow-up” has no time referent.
- Operator experience and centre volume should be pre-specified as subgroups. This ensures that a learning-curve explanation is a planned analysis rather than a post hoc explanation in the discussion section.
- Registries, regulatory files and surveillance reports should be added to the search. With dates, and with a stated rule on whether registry reports constitute included evidence or context.
- Reporting items should be counted. The proportion of included studies supplying each item should be published, as it is the review’s most reusable finding.
- The scope of the conclusion should be stated explicitly. Which generation, which setting, and an explicit statement of what a clinician cannot conclude from the review.
Frequently asked questions
Because the intervention is a composite rather than a dose. A device outcome depends on the model and generation implanted, the surgical approach and technique, the operator’s experience with that device, the implant position and sizing, and the perioperative and rehabilitation pathway, any of which can differ between two studies that name the same product. Three further differences are relevant: the device itself changes across generations while the literature accumulates, so a trade name may cover materially different objects; randomisation and blinding are often impossible, so comparative evidence is largely observational; and failures occur over years, in populations with substantial other-cause mortality, which makes simple proportions misleading and competing risks relevant.
Manufacturer, model, generation and catalogue or reference identifiers should be extracted for every included study, together with the implantation date range, which is usually the only available proxy for generation when model detail is missing. Generation should then be made an explicit part of the synthesis: studies should be grouped by generation where the data allow, generation should be pre-specified as a subgroup where sufficient studies permit, and the number of included studies that allowed a generation to be identified should be reported. Finally, the scope of the conclusion should be stated. A sentence of the form “these findings relate to implants placed between the stated years and may not apply to the current generation” is what makes the review usable, and pooling across fifteen years of a product line without it produces an average over a product history that may describe no currently available implant.
The decision should be made in the protocol and stated, because all three positions are defensible. National and regional device registries often hold the longest follow-up and the largest numbers available for an implant, sometimes broken down by brand and generation, so excluding them omits the best long-term evidence; including them introduces a different kind of data source, with a different case mix and different outcome ascertainment, into a review of published studies. The three options are to include registry reports as eligible evidence, to use them as contextual evidence reported separately from the synthesis, or to exclude them and state this. It is not defensible to search journals only and present the result as the totality of the evidence on long-term device survivorship.
Usually for three concurrent reasons. The definitions differ: revision, reoperation, clinical failure, radiographic failure and explant are not the same event, and studies draw the boundaries differently. The follow-up differs: a proportion at two years and one at ten years are different quantities, and pooling them as proportions discards the time structure entirely. The denominators are affected by competing mortality, which is substantial in an older implant population; a patient who dies before revision does not provide evidence of implant success in any useful sense. Where the definitions, horizons, generations and care contexts are genuinely compatible, standard meta-analysis applies. Where they are not, survivorship should be reported at stated timepoints with numbers at risk, together with a per-study table that keeps each figure alongside its definition, follow-up and generation.
It should be extracted, pre-specified, and its absence reported. The relevant fields are operator experience with the specific device, centre procedural volume, whether the study period covered the introduction of the device at that centre, whether any author was a developer of the device or practised in a designated centre of excellence, and whether proctoring or training was described. Operator experience and centre volume should be declared as pre-specified subgroups in the protocol, so that a learning-curve explanation is a planned analysis rather than a post hoc account of an unexpected result in the discussion. Most included studies will report none of these items; reporting the proportion that did is both an accurate statement about the evidence base and the reason why the pooled estimate should be interpreted cautiously.
For signal detection and history, yes; for incidence, no. Spontaneous and mandatory adverse event reporting databases have no denominator: they record how many events were reported, not how many devices were implanted or what proportion of events were reported, so no rate can be derived from them. They are genuinely useful for identifying failure modes that the published literature has not described, and for establishing the regulatory history of a device, including field safety notices, recalls and design changes following adverse outcomes. That history should be reported in a review, because it indicates which included studies describe the implant a reader is considering. The review should state explicitly that these sources were used for signal detection and context rather than for rate estimation.
An instrument designed for non-randomised studies of interventions should be used and named with its version, and any randomised evidence should be appraised separately with a randomised-trial instrument rather than the two appraisals being merged. The domains that carry most weight in device evidence are confounding, particularly by indication, patient selection and operator; selection into the study, including whether the series was consecutive; classification of the intervention, which for devices means whether model and generation were documented; deviations from the intended intervention, including intraoperative conversion to another device; missing data, which over the time horizon of an implant means loss to follow-up and whether those lost differ systematically; and outcome measurement, which for radiographic or clinician-assessed failure is assessor-dependent. Two device-specific points should be added: single-arm series cannot support comparative claims, and developer or designated-centre series are not estimates of general practice.
At least four limitations, stated explicitly rather than left to inference. A review usually cannot indicate how the current generation of the device will perform, because the evidence describes earlier generations. It usually cannot indicate how the device will perform in a particular clinician’s hands at a particular centre, because operator experience and centre volume are rarely reported and the included series often derive from adoption phases or specialist centres. It cannot convert a pooled proportion into an expected failure rate for an individual patient when follow-up horizons and competing mortality differ across studies. Nor can it establish comparative superiority over another device from separately reported single-arm series, however consistent they appear. A review that states these limitations is more useful than one that allows a pooled revision rate to be read as though it addressed all four.
References
- Chapter 4: Searching for and selecting studies. Cochrane Handbook for Systematic Reviews of Interventions, version 6.5. Cochrane. 2024;6.5. read 9 October 2026. cochrane.org
- Chapter 5: Collecting data. Cochrane Handbook for Systematic Reviews of Interventions, version 6.5. Cochrane. 2024;6.5. read 9 October 2026. cochrane.org
- Higgins JPT, Thomas J, Chandler J, Cumpston M, Li T, Page MJ, Welch VA (editors). Cochrane Handbook for Systematic Reviews of Interventions, version 6.5. Cochrane. 2024;6.5. read 9 October 2026. cochrane.org
- Metwally M, Raybould G, Cheong YC, Horne AW. Surgical treatment of fibroids for subfertility. Cochrane Database of Systematic Reviews. 2020;2020(1):CD003857. doi:10.1002/14651858.CD003857.pub4 read 9 October 2026. doi.org/10.1002/14651858.CD003857.pub4
- Moher D, Shamseer L, Clarke M, Ghersi D, Liberati A, Petticrew M, et al. Preferred reporting items for systematic review and meta-analysis protocols (PRISMA-P) 2015 statement. Systematic Reviews. 2015;4(1):1. doi:10.1186/2046-4053-4-1 read 9 October 2026. doi.org/10.1186/2046-4053-4-1
- Page MJ, McKenzie JE, Kirkham J, Dwan K, Kramer S, Green S, Forbes A. Bias due to selective inclusion and reporting of outcomes and analyses in systematic reviews of randomised trials of healthcare interventions. Cochrane Database of Systematic Reviews. 2014;10:MR000035. read 9 October 2026. cochrane.org
- Lipp A, Shaw C, Glavind K. Mechanical devices for urinary incontinence in women. Cochrane Database of Systematic Reviews. 2014;2014(12):CD001756. doi:10.1002/14651858.CD001756.pub6 read 9 October 2026. doi.org/10.1002/14651858.CD001756.pub6
- Ip S, Hadar N, Keefe S, Parkin C, Iovin R, Balk EM, Lau J. A web-based archive of systematic review data. Systematic Reviews. 2012;1:15. read 9 October 2026. cochrane.org
- Horsley T, Dingwall O, Sampson M. Checking reference lists to find additional studies for systematic reviews. Cochrane Database of Systematic Reviews. 2011;2011(8):MR000026. doi:10.1002/14651858.MR000026.pub2 read 9 October 2026. doi.org/10.1002/14651858.MR000026.pub2
- Manu M, Anand G. A review of medical device regulations in India, comparison with European Union and way-ahead. Perspectives in Clinical Research. 2022;13(1):3-11. doi:10.4103/picr.picr_222_20 read 9 October 2026. doi.org/10.4103/picr.picr_222_20
- Page MJ, McKenzie JE, Bossuyt PM, Boutron I, Hoffmann TC, Mulrow CD, et al. The PRISMA 2020 statement: an updated guideline for reporting systematic reviews. BMJ. 2021;372:n71. doi:10.1136/bmj.n71 read 9 October 2026. doi.org/10.1136/bmj.n71
- Bramer WM, de Jonge GB, Rethlefsen ML, Mast F, Kleijnen J. A systematic approach to searching: an efficient and complete method to develop literature searches. Journal of the Medical Library Association. 2018;106(4):531-541. doi:10.5195/jmla.2018.283 read 9 October 2026. doi.org/10.5195/jmla.2018.283
- Tricco AC, Lillie E, Zarin W, O'Brien KK, Colquhoun H, Levac D, et al. PRISMA Extension for Scoping Reviews (PRISMA-ScR): Checklist and Explanation. Annals of Internal Medicine. 2018;169(7):467-473. doi:10.7326/M18-0850 read 9 October 2026. doi.org/10.7326/M18-0850
- Mayo-Wilson E, Li T, Fusco N, Dickersin K. Practical guidance for using multiple data sources in systematic reviews and meta-analyses (with examples from the MUDS study). Research Synthesis Methods. 2018;9:2-12. read 9 October 2026. cochrane.org
- Page MJ, Altman DG, Shamseer L, McKenzie JE, Ahmadzai N, et al. Reproducible research practices are underused in systematic reviews of biomedical interventions. Journal of Clinical Epidemiology. 2018;94:8-18. doi:10.1016/j.jclinepi.2017.10.017 read 9 October 2026. doi.org/10.1016/j.jclinepi.2017.10.017
- Munn Z, Peters MD, Stern C, Tufanaru C, McArthur A, Aromataris E. Systematic review or scoping review? Guidance for authors when choosing between a systematic or scoping review approach. BMC Medical Research Methodology. 2018;18(1):143. doi:10.1186/s12874-018-0611-x read 9 October 2026. doi.org/10.1186/s12874-018-0611-x
- Mayo-Wilson E, Li TJ, Fusco N, Bertizzolo L, Canner JK, Cowley T, et al. Cherry-picking by trialists and meta-analysts can drive conclusions about intervention efficacy. Journal of Clinical Epidemiology. 2017;91:95-110. read 9 October 2026. cochrane.org
- Baudard M, Yavchitz A, Ravaud P, Perrodeau E, Boutron I. Impact of searching clinical trial registries in systematic reviews of pharmaceutical treatments: methodological systematic review and reanalysis of meta-analyses. BMJ. 2017;356:j448. read 9 October 2026. cochrane.org
- Bramer WM, Giustini D, de Jonge GB, Holland L, Bekhuis T. De-duplication of database search results for systematic reviews in EndNote. Journal of the Medical Library Association. 2016;104(3):240-243. doi:10.3163/1536-5050.104.3.014 read 9 October 2026. doi.org/10.3163/1536-5050.104.3.014
- Cao AM, Cox MR, Eslick GD. Study design in evidence-based surgery: What is the role of case-control studies? World Journal of Methodology. 2016;6(1):101-104. doi:10.5662/wjm.v6.i1.101 read 9 October 2026. doi.org/10.5662/wjm.v6.i1.101
- Jones CW, Keil LG, Holland WC, Caughey MC, Platts-Mills TF. Comparison of registered and published outcomes in randomized controlled trials: a systematic review. BMC Medicine. 2015;13:282. read 9 October 2026. cochrane.org
- Rethlefsen ML, Farrell AM, Osterhaus Trzasko LC, Brigham TJ. Librarian co-authors correlated with higher quality reported search strategies in general internal medicine systematic reviews. Journal of Clinical Epidemiology. 2015;68(6):617-626. doi:10.1016/j.jclinepi.2014.11.025 read 9 October 2026. doi.org/10.1016/j.jclinepi.2014.11.025
- Stewart LA, Clarke M, Rovers M, Riley RD, Simmonds M, Stewart G, Tierney JF; PRISMA-IPD Development Group. Preferred reporting items for a systematic review and meta-analysis of individual participant data: the PRISMA-IPD statement. JAMA. 2015;313:1657-1665. read 9 October 2026. cochrane.org
- Hoffmann TC, Glasziou PP, Boutron I, Milne R, Perera R, Moher D, et al. Better reporting of interventions: template for intervention description and replication (TIDieR) checklist and guide. BMJ. 2014;348:g1687. read 9 October 2026. cochrane.org
- Turner EH. How to access and process FDA drug approval packages for use in research. BMJ. 2013;347. read 9 October 2026. cochrane.org
- Zarin DA, Tse T, Williams RJ, Califf RM, Ide NC. The ClinicalTrials.gov results database: update and key issues. New England Journal of Medicine. 2011;364:852-860. read 9 October 2026. cochrane.org
- Papaioannou D, Sutton A, Carroll C, Booth A, Wong R. Literature searching for social science systematic reviews: consideration of a range of search techniques. Health Information and Libraries Journal. 2010;27(2):114-122. doi:10.1111/j.1471-1842.2009.00863.x read 9 October 2026. doi.org/10.1111/j.1471-1842.2009.00863.x
- Farrokhyar F, Karanicolas PJ, Thoma A, Simunovic M, Bhandari M, Devereaux PJ, et al. Randomized controlled trials of surgical interventions. Annals of Surgery. 2010;251(3):409-416. doi:10.1097/SLA.0b013e3181cf863d read 9 October 2026. doi.org/10.1097/SLA.0b013e3181cf863d
- Grant MJ, Booth A. A typology of reviews: an analysis of 14 review types and associated methodologies. Health Information and Libraries Journal. 2009;26(2):91-108. doi:10.1111/j.1471-1842.2009.00848.x read 9 October 2026. doi.org/10.1111/j.1471-1842.2009.00848.x
- Lederle FA, Freischlag JA, Kyriakides TC, Padberg FT, Matsumura JS, Kohler TR, et al. Outcomes following endovascular vs open repair of abdominal aortic aneurysm: a randomized trial. JAMA. 2009;302(14):1535-1542. doi:10.1001/jama.2009.1426 read 9 October 2026. doi.org/10.1001/jama.2009.1426
- Gotzsche PC, Hrobjartsson A, Maric K, Tendal B. Data extraction errors in meta-analyses that use standardized mean differences. JAMA. 2007;298:430-437. read 9 October 2026. cochrane.org
- Buscemi N, Hartling L, Vandermeer B, Tjosvold L, Klassen TP. Single data extraction generated more errors than double data extraction in systematic reviews. Journal of Clinical Epidemiology. 2006;59:697-703. read 9 October 2026. cochrane.org
- Jones AP, Remmington T, Williamson PR, Ashby D, Smyth RL. High prevalence but low impact of data extraction and reporting errors were found in Cochrane systematic reviews. Journal of Clinical Epidemiology. 2005;58:741-742. read 9 October 2026. cochrane.org
Every source above is a peer-reviewed article, a publisher’s record of one, or a standard issued by the body that owns it. Each carries the date it was last read. Where a standard moves, this article is revised rather than left up with a note.
How to cite this article
Pick your referencing style and copy it. Your university or target journal may vary the punctuation — check its guide if it has one.
Pubrica Academy. (2026, October 9). Reporting Quality in Systematic Reviews of Implantable Devices. Pubrica. https://pubrica.com/academy/systematic-review/systemic-reviews-on-implantable-medical-devices-provide-a-quality-of-reporting/
Pubrica Academy. “Reporting Quality in Systematic Reviews of Implantable Devices.” Pubrica, 9 Oct. 2026, pubrica.com/academy/systematic-review/systemic-reviews-on-implantable-medical-devices-provide-a-quality-of-reporting/.
Pubrica Academy (2026) Reporting Quality in Systematic Reviews of Implantable Devices. Pubrica. Available at: https://pubrica.com/academy/systematic-review/systemic-reviews-on-implantable-medical-devices-provide-a-quality-of-reporting/ (Accessed: 15 October 2026).
Pubrica Academy. “Reporting Quality in Systematic Reviews of Implantable Devices.” Pubrica. October 9, 2026. https://pubrica.com/academy/systematic-review/systemic-reviews-on-implantable-medical-devices-provide-a-quality-of-reporting/
Pubrica Academy, ‘Reporting Quality in Systematic Reviews of Implantable Devices’ (Pubrica, 9 October 2026) <https://pubrica.com/academy/systematic-review/systemic-reviews-on-implantable-medical-devices-provide-a-quality-of-reporting/> accessed 15 October 2026.
Pubrica Academy. Reporting Quality in Systematic Reviews of Implantable Devices. [Internet]. Pubrica; 2026 Oct 9 [cited 2026 Oct 15]. Available from: https://pubrica.com/academy/systematic-review/systemic-reviews-on-implantable-medical-devices-provide-a-quality-of-reporting/
Pubrica Academy, “Reporting Quality in Systematic Reviews of Implantable Devices,” Pubrica, Oct. 9, 2026. [Online]. Available: https://pubrica.com/academy/systematic-review/systemic-reviews-on-implantable-medical-devices-provide-a-quality-of-reporting/
@misc{systemic_reviews_on_implantable_medica_2026,
author = {{Pubrica Academy}},
title = {Reporting Quality in Systematic Reviews of Implantable Devices},
howpublished = {Pubrica},
year = {2026},
month = {oct},
url = {https://pubrica.com/academy/systematic-review/systemic-reviews-on-implantable-medical-devices-provide-a-quality-of-reporting/},
note = {Accessed: <span data-cite-acc="iso">2026-10-15</span>}
}This article is written to be learned from. Cite it if you are referring to what it says; if you are using the evidence behind it, cite the sources listed above instead.
Extraction and reporting checklist for an implantable device review
The per-study block should be used during extraction and the review-level block during drafting. The proportion-reported column at the end is the output that other researchers are most likely to cite.
IMPLANTABLE DEVICE REVIEW — EXTRACTION AND REPORTING
For every field: R = reported, N = not reported. Do NOT infer.
=== PER STUDY: THE DEVICE ===
Manufacturer: ________________________________ [R/N]
Model / product name: ________________________ [R/N]
GENERATION or version: _______________________ [R/N]
Catalogue / reference identifiers: ___________ [R/N]
Implantation date range: _____________________ [R/N]
(use as generation proxy where model detail is absent)
Known field safety notice / recall / design change
affecting this device or period? [ ] yes: ______ [ ] no
=== PER STUDY: THE PROCEDURE ===
Surgical approach: ___________________________ [R/N]
Fixation / attachment method: ________________ [R/N]
Implant position / sizing / configuration: ___ [R/N]
Adjuncts or additional components used: ______ [R/N]
Perioperative protocol: ______________________ [R/N]
Rehabilitation / follow-up protocol: _________ [R/N]
Intraoperative conversion to another device? [R/N] ____
=== PER STUDY: THE OPERATOR AND SETTING ===
Operator experience with THIS device: ________ [R/N]
Centre procedural volume: ____________________ [R/N]
Did the study period include the device's introduction
at that centre? [ ] yes [ ] no [ ] cannot tell
Any author a DEVELOPER of the device? [ ] yes [ ] no
Designated centre of excellence / reference centre? [ ] yes [ ] no
Proctoring or training programme described? [R/N]
Country / health system: _____________________ [R/N]
=== PER STUDY: OUTCOMES (verbatim, do not map yet) ===
Outcome 1 name as used by the study: _________________
DEFINITION, QUOTED VERBATIM: ______________________
Which category does it correspond to?
[ ] revision [ ] reoperation [ ] clinical failure
[ ] radiographic failure [ ] explant [ ] survivorship
[ ] cannot map — record and exclude from that synthesis
Events: ____ Denominator: ____
Outcome 2 ... (repeat)
FOLLOW-UP
Minimum: ____ Median: ____ Maximum: ____ [R/N]
Number at risk at each reported timepoint: ____ [R/N]
Loss to follow-up: ____ Characteristics of those lost: [R/N]
Competing mortality reported? [R/N] How handled: ______
=== PER STUDY: FUNDING AND INVOLVEMENT ===
Funding source: ______________________________ [R/N]
Manufacturer involvement in:
[ ] funding [ ] design [ ] conduct [ ] analysis
[ ] authorship [ ] device provision [ ] none stated
Declared conflicts: __________________________ [R/N]
=== PER STUDY: APPRAISAL ===
Design: [ ] RCT [ ] non-randomised comparative [ ] single-arm series
Instrument used + VERSION: ___________________________
(appraise RCTs and non-randomised studies SEPARATELY)
Confounding by indication / selection / operator: ______
Consecutive series? [ ] yes [ ] no [ ] cannot tell
Intervention classification (model/generation documented): ______
Outcome assessor independent? [R/N]
=== REVIEW LEVEL ===
Sources searched beyond journals:
[ ] device registries — named: ____________ date ______
Role: [ ] included evidence [ ] context only [ ] excluded
[ ] regulatory submission / approval documents: ______
[ ] summaries of safety and clinical performance: ______
[ ] post-market surveillance / follow-up reports: ______
[ ] adverse event databases — FOR SIGNAL ONLY, no rates [ ]
Generation handling in the synthesis: ________________
Operator experience pre-specified as a subgroup? [ ] yes [ ] no
Competing risks addressed? [ ] yes, how: ______ [ ] no
Outcome-definition MAPPING TABLE published? [ ] yes
Distribution of manufacturer involvement reported? [ ] yes
=== PROPORTION-REPORTED TABLE (publish this) ===
Item | Studies reporting it | % of included studies
Device generation | ____ | ____
Implantation date range | ____ | ____
Surgical technique | ____ | ____
Operator experience | ____ | ____
Centre volume | ____ | ____
Verbatim outcome definition | ____ | ____
Numbers at risk | ____ | ____
Manufacturer involvement | ____ | ____
=== SCOPE OF CONCLUSIONS (write these sentences) ===
"These findings relate to device generation(s) ______,
implanted between ______ and ______, in ______ settings."
"This review cannot establish ____________________________."
Generation could not be determined for ____ of ____ studies. [ ] stated Pubrica AcademyWritten and maintained by Pubrica’s subject matter experts. Articles are revised when the underlying standards change, and each article states the date on which it was last reviewed.
How Pubrica can help
Pubrica builds device review extraction around the items that determine transferability (model and generation, implantation dates, technique, operator experience, verbatim outcome definitions and numbers at risk) and searches registries, regulatory documents and post-market surveillance alongside the journal literature. The deliverables comprise the outcome-definition mapping table, the proportion-reported table and a conclusion explicitly scoped to the generations and settings that the evidence covers.