Reporting of morbidity data in general surgery: a one-year cross-sectional analysis of the six journals with the highest impact factor

Abstract

Background:

Morbidity is a critical outcome in general surgery studies, yet its reporting often lacks methodological transparency. The objective of this cross-sectional study is to describe how postoperative morbidity is reported in high-impact general surgery journals and to propose key elements that could serve as a basis for future reporting standards.

Materials and methods:

All articles reporting postoperative complications and published online throughout 2022 in the following journals were selected: JAMA Surgery, Annals of Surgery, International Journal of Surgery, British Journal of Surgery, World Journal of Emergency Surgery and Journal of American College of Surgeons. Quality standards were established, forming the variables in the morbidity data analysis. A descriptive analysis and a probability model were completed to assess the applicability of these standards.

Results:

249 studies were included. Consistent information about morbidity sources was not provided in 247 studies (99.1%). Those responsible for registering complications were not identified in 207 (83.1%) studies. Follow-up was conducted for patients in 153 (61%) studies, but in only 67 (26.9%) cases methods were discussed. A standardized classification was applied in 157 (63%) studies, with the Clavien–Dindo classification being the most used score (n = 140, 49.8%). Among these, 16 different variations of the classification were identified. A specific period of morbidity data collection was not determined in 104 (41.8%) studies.

Conclusion:

There is no standardized methodology to register and report morbidity in high impact surgery publications. Disparity and subjectivity in reporting results translate into bias and limited comparability. Authors and readers should aim for a more standardized and transparent approach to morbidity.

Introduction

Postoperative complications (PC) represent the most valuable quality indicators in surgical procedures. While PC are normally reported in General Surgery studies, the methodology used to register, collect, analyze, and follow them is often omitted.

This lack of information can obscure bias, which could significantly impact the validity of outcomes in surgical studies, including those regarding PC. These biases are crucial, particularly when applied to the validation of new surgical techniques.

HIGHLIGHTS

• Morbidity is crucial in General Surgery studies but often lacks standardized reporting.

• This study analyzed 249 articles from six major surgery journals published in a year.

• 99.1% of studies did not consistently report morbidity sources.

• 61% had patient follow-up; only 26.9% discussed methods.

• 63% used standardized classification, with Clavien–Dindo being the most common; 16 variations of this classification were used.

• Conclusion: Need for standardized and transparent morbidity reporting method to reduce bias and improve comparability.

Without a uniform classification and reporting system, complications may be underreported, misclassified, or omitted, resulting in an inaccurate assessment of actual morbidity and a lack of transparency^[1].

Among the many consequences, the inability to determine the true morbidity in surgery stands out. It follows that morbidity related to similar procedures cannot be compared, a gold standard cannot be established, and new surgical techniques cannot be validated.

Beyond research limitations, the failure to implement standardized morbidity reporting has broader implications for patient safety and healthcare policy. The absence of transparent, audited complication records can lead to suboptimal decision-making regarding the centralization of complex procedures, surgeon training, and resource allocation, hindering the ability to benchmark surgical services effectively and ultimately impacting the quality of care and economic efficiency of surgical services.^[2]

Tools to report postoperative morbidity in a unified and objective manner are available, namely, the Clavien–Dindo classification (CDC)^[3] and the Comprehensive Complication Index^[4]. Both have been validated clinically and economically, over a 90-days period.^[2,5]

Previous studies have identified the lack of standardized methods and introduced the problem deriving from this, as well as several criteria to consider when evaluating how complications are reported in a General Surgery department.^[1,6]

In the light of this, there are few examples of efforts being made in this direction to standardize publication of PC.^[7]

Other publications have introduced the analysis of PC’s collecting methods. They either provided an overview on a national level, outlining the disparity in the methodology within one country^[8], or focused on data sources and how combining medical and nursing notes increases the quality of data collected on PC^[9].

Nevertheless, to date, no reviews exist on how morbidity is determined in the field of General Surgery, nor is there a standardized methodology that researchers can rely on.

The study was based on the hypothesis that high-impact papers use a fully developed and standardized methodology to record PC, thereby adhering to basic quality standards.

The main research question addressed was: To what extent do high-impact General Surgery journals apply standardized methods when reporting morbidity?

The objective was to fill the research gap by outlining the methods behind publishing morbidity in high impact General Surgery studies. The secondary objective was to select which elements represented the basis of these results so as to establish them as quality criteria and acceptable standards. These elements meant to represent different facets of morbidity in surgery and included variables from the patient’s timeline around a surgical procedure.

Material and methods

Study design

A cross-sectional study was designed, choosing the number of journals included based on the articles published monthly, and aiming for a representative number.

This article was prepared in accordance with the STROBE guidelines and its clarifications^[10,11], and reported in line with the STROCSS criteria^[12].

The journal impact factor (JIF) was used to select the publications for this study. The articles published in these journals are universally expected to be of the best quality, as they fulfill strict criteria before publication. This criterion was used to ensure a valid representation of the best scientific publications available to professionals.

As a first measure to avoid selection bias, any journal specializing in a specific field of General Surgery was excluded.

Under the term “General Surgery”, the following fields were included: abdominal surgery (including colorectal, upper gastrointestinal, bariatric, endocrine, abdominal wall, oncologic, hepatopancreaticobiliary, minimally invasive surgeries and transplant); trauma and breast surgeries.

Six journals were included in this analysis, chosen based on their JIF for the year 2021 among General Surgery publications^[13]. The following were selected: JAMA Surgery (2021 JIF: 16.685), Annals of Surgery (2021 JIF: 13.787), International Journal of Surgery (2021 JIF: 13.400), British Journal of Surgery (2021 JIF: 11.782), World Journal of Emergency Surgery (2021 JIF: 8.165) and Journal of American College of Surgeons (2021 JIF: 6.532).

Two researchers performed a two-step review of all the articles published in each selected journal from January to December 2022.

The one-year time frame was intentionally selected to capture the most recent and up-to-date literature, reflecting current morbidity reporting practices in high-impact journals while minimizing variability over time.

The initial screening was conducted by reviewing titles and abstracts, identifying a set of articles composed of scientific studies (clinical trials, cohort studies, case-control studies and, cross-sectional studies) whose topic included General Surgery-related data.

The second selection phase applied predefined exclusion criteria. Articles were excluded if they were unrelated to General Surgery or classified as viewpoints, editorials, letters, commentaries, guidelines, meta-analyses, reviews, consensus statements, case reports, surveys, or corrections/errata.

Further exclusions applied to studies that lacked data on surgery-related complications, did not involve a surgical procedure, or exclusively reported mortality. Additionally, studies explicitly stating that no complications were recorded were also excluded.

This step was performed by reviewing the full content of the article, combined with a search by keyword to avoid any omission errors.

The keywords (alone and in combination) used for the mentioned search were: “complications,” “morbidity,” “follow-up,” “readmission,” “bias,” “informed + consent,” “competing + interest.”

For more details about the exclusion criteria applied, see diagram in Figure 1.

Figure 1.

Diagram of the selection process. AS: Annals of Surgery; IJS: International Journal of Surgery; BJS: British Journal of Surgery; WJES: World Journal of Emergency Surgery; JACS: Journal of American College of Surgeons.

Variables and data collection

Fifteen variables were chosen to include the largest amount of data possible, taking into consideration the expected diversity of content across research settings. This decision was based on criteria included in previous literature, a preliminary sample analysis, and structured brainstorming sessions.^[2,6]

These variables reflect key aspects of morbidity reporting methodology commonly emphasized in the literature and were selected to capture the structural elements most relevant to transparency and reproducibility.

It is not possible to reach comparable and transparent morbidity results without including detailed information about methodology.

Therefore, reporting morbidity should involve a precise presentation of the sources consulted, the time of collection, and the classifications used. It is also necessary to include data about the post-discharge period (follow-up, time and method, readmissions, and extra-hospital complications), along with defining who completes the task and what relationships they have with patients (this also implies any bias or competing interest).

Considering these criteria as the minimum standard for a comparable method for morbidity publication, it was analyzed how they were fulfilled in the articles included.

A brief explanation and specific considerations of each variable are provided below.

The study type was determined based on what was declared by the authors. If not specified, it was categorized according to the design of the study.

All studies that included General Surgery-related procedures were accepted. Studies that included procedures which are generally performed by different specialists, were taken into consideration only if a general surgeon participated to the surgery. vascular and pediatric surgeries were not included.

If specified, the period of morbidity collection was registered, in terms of how many postoperative days were considered for a complication to happen and to be registered as part of the morbidity results.

Sources of morbidity data were analyzed, focusing on details that could clarify where and by whom those data were collected. Specific details were considered, such as who recorded the complications (doctors, nurses, or other healthcare professionals) and whether the records came from hospital notes, a morbidity recording system, test results, or other written documentation.

Who conducted the data collection was also revised. Any data collector that was stated as responsible for the task was registered, whether they were part of the research team or other health care professionals not involved in the investigation.

Assessed comorbidity was documented. The focus was on standardized classifications, scales, and scores, like the ASA score (American Society of Anaesthesiologists Physical Status Classification System) or the Charlson Comorbidity Index. The comorbidity that was specific to the subject of the study (as the ECOG performance status, the SOFA score, etc.) was classified as an independent outcome.

The method of morbidity classification was registered by categorizing all results about postoperative complications by the chosen method of presentation.

The first approach was dividing results that employed a classification of morbidity from those that did not.

All morbidity selected without any specified criteria, as well as any specific complication related to the topic of the study, were considered as unclassified morbidity.

Additionally, among the results that included a classification, all scales and scores used were detailed, alone or in combination with other classifications or with unclassified morbidity.

A different approach was necessary to assess the use of the Clavien–Dindo Classification (CDC). This scale has a broad range of application, and we observed that it was often adapted to the specific outcomes of each study. For the purpose of this analysis, we considered the correct use to be the unmodified application of the CDC, including all of its original grades (I to V). Any variation, such as excluding certain grades, aggregating complications into broader categories (e.g., major/minor), or using the scale selectively, has been considered a misuse.

It was also noted if intraoperative morbidity was included as part of perioperative complications.

Follow-up was first recorded as a binary variable, listing if it was carried out or not, then the chosen method to execute it was analyzed.

From a similar perspective, readmissions and complications that occurred in an extra-hospital setting were registered, together with any documented informed consent offered to patients.

Lastly, any bias and competing interests reported by the authors were analyzed. The perspective chosen was to find references to morbidity, so biases were categorized as related and non-related to it, registering if any competing interests were reported as well.

Specific references to patients treated by the researchers were sought, emphasizing that direct involvement in recording morbidity should be considered and declared as a competing interest.

Six of these variables (time of morbidity collection, data sources, data collectors, morbidity classification methods, follow-up, and readmissions) were selected as core variables based on their emphasis in the literature and their relevance to the structural aspects of morbidity reporting. Their selection resulted from a focused review of methodological literature and iterative discussions within the working group, prioritizing factors shown to influence transparency, reproducibility, and comparability in postoperative outcomes research. These variables are not intended as a formal quality index but serve as a descriptive framework to analyze current reporting practices.

Bias addressment

Data collection was performed by two members of the research team. In case of a discordance of opinion, a third investigator was involved.

Other measures to avoid potential bias were taken in this process, using the combined research method previously described.

The comorbidity assessment showed a notable disparity in these data, making it difficult to compare them. To some extent, this is due to the lack of classifications with a range of features broad enough so to allow exclusive use.

To be able to compare comorbidity results, we focused on data expressed using classifications to mitigate any bias encountered. Yet, we believe that the potential value of these data was partially lost due to their heterogeneity.

Statistical analysis

Data were imported into the R Statistical Software (v 4.3.1) and all variables were processed to design a descriptive analysis. Graphics were created using the ggplot R package.

A weighted kappa coefficient was calculated for each variable based on a random sample of 30 articles independently reviewed by two observers. A global weighted kappa and a corresponding z statistic were also calculated to test agreement beyond chance.

The multinomial probability model was designed using simplified data from the six main variables, defining seven outputs (scenarios). Each scenario marked the probability of finding an article which provided data for any combination of the six main variables, ordered from zero out of six, to six out of six. To achieve this, specific parameters for a multinomial probability distribution were later estimated from the observations of 2022. A 95% confidence interval (CI) was applied.

A Poisson distribution function was used to model the sampling behavior of the data. Maximum likelihood estimators were applied for parameter estimation. To calculate the Bayes factor and the posterior probability of the theoretical optimistic model, a Bayesian inferential approach was employed, using conjugate Poisson-normal and Poisson-gamma distributions.

Results

Review and selection process

During the 12 months of 2022, 2303 articles were published in the six journals reviewed, and a total of 18 068 months’ worth of data was considered for this study.

Two hundred and eighteen studies (87%) were exclusively General Surgery-related, while twenty-nine included surgeries from other specialties in their results.

The selection process, the exclusion criteria, and the final selection of articles are detailed in Figure 1.

References to all articles included are listed in the Supplementary Digital Content Material, available at: http://links.lww.com/JS9/E952.

Data sources and morbidity data collectors

A total of 99.1% (n = 247) of the studies did not provide consistent information about the sources used to collect perioperative complications. Among them, 40.5% gave no information, while 58.5% cited the electronic clinical record or a database as their source, without clarifying what was reviewed to obtain the data. Databases with standardized records are included in these numbers, as none clarified their internal sources. Only two articles (both cross-sectional studies) specified all sources used, detailing which parts of the clinical record were taken into consideration. Results regarding data sources are shown in Table 1.

Table 1.

Data sources results, enumerated by journal

Data sources	JAMA	AS	IJS	BJS	WSES	JACS	Total
Not specified	10	40	22	17	3	10	n = 101 (40.5%)
Clinical record (no details)	4	24	8	7	6	6	n = 55 (22%)
Database (no details)							n = 80 (32.1%)
Multicenter database	8	26	4	10	2	13	63
Single center database	0	6	4	1	0	6	17
Clinical record + database (no details)							n = 3
Clinical record + multicenter database (no details)	1	0	0	0	0	0	1
Clinical record + single center database (no details)	0	1	0	0	0	1	2
Detailed records							2
Clinical record (entries from doctors, nurses, pharmacists, nutritionists, rehabilitators; medication records; radiological studies)	0	1	0	0	0	0	1
Multicenter database (hospital discharge abstracts; ambulatory care register; emergency room visits reports)	0	1	0	0	0	0	1
Others							n = 8
Intraoperative notes (anesthesia) + notes from outpatients’ clinic	0	0	0	0	0	1	1
Intraoperative clinical record (anesthesia)	0	0	0	0	0	1	1
Clinical interview	0	0	1	0	0	0	1
Clinical record (no details) + clinical interview	1	1	1	1	0	1	5

JAMA: Journal of the American Medical Association, surgery section; AS: Annals of Surgery; IJS: International Journal of Surgery; BJS: British Journal of Surgery; WJES: World Journal of Emergency Surgery; JACS: Journal of American College of Surgeons.

Who completed the data collection task was not reported in 83.1% (n = 207) of the studies. Only 12 studies mentioned the identity of the people involved (nurses, doctors, or both). The remaining articles generally referred to members of the investigation team. Specific results are shown in Table S1 of the Supplementary Digital Content Material, available at: http://links.lww.com/JS9/E952.

Follow-up and readmissions

Patients received follow-up in 61% (n = 153) of the studies, but in only 67 (26,9%) were the methods discussed. None of the articles explicitly stated that morbidity was registered during follow-up. Complete follow-up data are shown in Table S2 of the Supplementary Digital Content Material, available at: http://links.lww.com/JS9/E952.

Readmissions were reported in 46.9%(n = 117) of the studies, while 132 articles (53.1%) did not include them.

Morbidity classification

Overall, 35.3% (n = 88) of the articles did not use any classification for their morbidity results. Standardized classifications were applied in 63% (n = 157) of the studies, with the CDC being the most frequently used score. Other classifications, such as the Comprehensive Complication Index, were used far less frequently (6.8%). Notably, complications were not always treated as a multinomial variable; 22% (n = 55) of the studies reported them as binomial. Among those that used the CDC, there was considerable variability in its application, resulting in 16 different variations.

In 37 studies the CDC was used without modifications, including all its grades (I to V); 33 studies applied modified versions while still maintaining all grades. The remaining 54 articles (43.6%) used the CDC partially – most often omitting minor complications and reporting only grade III or higher. Details are presented in Table 2.

Table 2.

Morbidity classification data

	JAMA	Annals	IJS	BJS	WSES	JACS	Total
Not specified	1	3	0	0	0	0	4
Clavien–Dindo classification							n = 124 (49.8%)
CDC I–V	1	15	10	6	1	4	37
CDC I–V + SM	1	4	2	3	1	1	12
CDC I–V (yes/no) + SM	0	0	0	1	0	0	1
CDC major /minor	2	9	2	2	0	3	18
CDC major /minor + SM	0	1	1	0	0	0	2
CDC ≥ III	1	1	0	2	0	1	5
CDC ≥ III (yes/no) + SM	0	6	5	3	0	6	20
CDC ≥ IIIb	0	0	0	1	0	0	1
CDC ≥ II	0	0	0	1	0	0	1
CDC ≥ II + SM	0	1	0	1	0	0	2
CDC ≥ II (yes/no)	0	0	3	1	2	0	6
CDC IV–V	0	0	0	0	1	0	1
CDC ≤ III	0	0	1	1	0	0	2
CDC I/ ≥ II (yes/no)	0	0	1	0	0	0	1
Adapted Clavien–Dindo in Trauma Scoring System	0	1	0	0	0	0	1
Comprehensive Complication Index only	1	2	0	2	1	0	n = 6
Clavien–Dindo classification + Comprehensive Complication Index							n = 16
CDC major /minor + CCI	0	6	1	1	1	0	9
CDC ≥ III + CCI + SM	0	1	1	3	0	0	5
CDC ≥ III + CCI	0	2	0	0	0	0	2
Other classifications							11
National Surgical Quality Improvement Program	3	3	0	0	0	0	6
POSSUM criteria	0	0	2	0	0	0	2
Accordion Severity Grading System + Postoperative morbidities indexes + average complication burdens	0	1	0	0	0	0	1
MBSAQIP database	0	0	0	0	0	1	1
Veterans Affairs Surgical Quality Improvement	1	0	0	0	0	0	1
No classifications							n = 88 (35.3%)
SM	1	1	1	0	0	3	6
Non-specific complications	9	27	5	5	4	14	64
Non-specific complications + SM	0	4	1	0	0	0	5
Any complication (yes/no)	1	6	1	1	0	4	13

CDC: Clavien–Dindo Classification, followed by the interval of grades included (I to V); CCI: Comprehensive Complication Index; SM: Specific Morbidity, related to de subject studied. Non-specific complications: morbidity selected without specifying any criteria; MBSAQIP: Metabolic and Bariatric Surgery Accreditation and Quality Improvement Program. JAMA: Journal of the American Medical Association, surgery section; AS: Annals of Surgery; IJS: International Journal of Surgery; BJS: British Journal of Surgery; WJES: World Journal of Emergency Surgery; JACS: Journal of American College of Surgeons.

Intraoperative morbidity was included in 62 studies (24.8%), while extra-hospital morbidity was registered in 17 studies. Combining these data, 70.6% of the articles (n = 176) accepted as morbidity only complications that occurred in a post-operative and intra-hospital context.

Time of morbidity collection

A specific period of morbidity data collection was not defined in 104 (41.8%) of the studies.

When specified, the most common approach was collecting postoperative morbidity during a 30-day period (107 studies, 42.9%), followed by a 90-day period (24 studies). Fourteen studies used a time frame adapted to the design of the study.

Comorbidity

Comorbidity was frequently adapted to the subject of the study, presented through various factors deemed relevant by the authors, thereby causing a notable disparity in our results.

All studies included basic epidemiological information (at least age and sex).

As shown in Table 3, up to 56.6% (n = 141) of the studies used at least one classification, the most employed being the ASA classification, followed by the Charlson Comorbidity Index. However, a considerable number of articles did not use any standardized classification and only assessed specific morbidity (35.7%, n = 89). Lastly, nineteen articles did not provide comorbidity data.

Table 3.

Comorbidity results, listed by journal

Comorbidity	JAMA	Annals	IJS	BJS	WSES	JACS	Total
Not specified	0	2	0	2	0	2	6
Epidemiologic data only	3	5	1	0	0	4	13
SC only	8	40	14	9	6	12	n = 89 (35.7%)
ASA							n = 112 (44.9%)
ASA only	2	2	2	0	0	0	6
ASA + NSQIP	1	0	0	0	0	0	1
ASA + SC	8	42	15	19	3	17	105
Charlson Comorbidity Index							n = 11
Charlson Comorbidity Index only	2	0	1	0	0	2	5
Charlson Comorbidity Index + SC	0	2	1	2	0	1	6
Charlson Comorbidity Index + ASA	0	3	4	3	2	0	12
Charlson Comorbidity Index + ASA + SC	0	3	0	0	0	1	4
Others							2
SC + Elixhauser comorbidity index	0	1	0	0	0	0	1
SC + POSSUM criteria	0	0	1	0	0	0	1

SC: Specific Comorbidity, related to the subject studied; ASA: American Society of Anaesthesiologists Physical Status Classification System; NSQIP: National Surgical Quality Improvement Program (US). JAMA: Journal of the American Medical Association, surgery section; AS: Annals of Surgery; IJS: International Journal of Surgery; BJS: British Journal of Surgery; WJES: World Journal of Emergency Surgery; JACS: Journal of American College of Surgeons.

Morbidity-related bias, competing interests, and informed consent

Only 28 studies (11.2%) referred to any bias related to morbidity. In contrast, 100 studies (40.2%) did not mention any potential bias. Competing interests were declared in 27 studies, none of which referenced factors that might affect morbidity data. No article acknowledged potential biases stemming from investigators involved in patient care. Informed consent for study participation was reported in 71 studies (28.5%).

Cohen’s kappa

The global weighted kappa was 0.78. The z statistic derived from this value was 2.91, CI (z) = (2.39–3.44), P = 0.001, indicating statistically significant agreement between reviewers.

Probability model and gold standard applicability

A multinomial probability model was used on simplified data from six main variables to assess adherence to basic quality standards in morbidity data.

The optimal scenario would have been finding a study that included any information for all six variables. The probability of finding this scenario in this model was 2.8% (n = 7, CI = 0–5.9%).

Subsequently, similar scenarios were analyzed, with decreasing number of variables found (considering any combination of variables as valid). The following results have been found:

• Five out of six variables: 12.1% (n = 30, CI = 10.5–13.6%)

• Four out of six: 26.5% (n = 66, CI = 24.1–28.9%)

• Three out of six: 27.7% (n = 69, CI = 26.1–29.3%)

• Two out of six: 22.5% (n = 56, CI = 19.3–25.6%)

• One out of six: 7.6% (n = 19, CI = 4.5–10.8%)

• Zero out of six: 0.8% (n = 2, CI = 0–1.6%)

These results are shown in Figure 2.

Figure 2.

Probability model showing adherence to six variables set as basic standards of quality in morbidity data. Each bar represents a scenario (ranging from 0 to 6 variables present in the article). Black lines on top of each bar indicate the 95% confidence interval (CI) for the estimated proportion. The mean is represented by the gray vertical line.

A composite score from 0 to 6 was constructed based on the presence of the six core variables, representing the amount of methodological information reported in each article. The mean score across the sample was 3.2 (95% CI: 3.0–3.35). The probability of obtaining a score below 2 in our sample was 0.38%, while the probability of obtaining a score above 5 was 0.1%.

To evaluate whether the sample was statistically compatible with an optimistic distribution model (defined as an expected score with an interquartile range between 4 and 5, and a prior probability of 0.8), a Bayesian hypothesis test was performed. The resulting Bayes factor, comparing the optimistic model to a likelihood-based alternative, was 0.02, indicating statistical incompatibility between the optimistic model and the observed data.

To reinforce this conclusion, the posterior probability that our sample was generated under the optimistic model was also calculated, yielding a value of 0.09, further supporting the lack of compatibility between the observed data and the optimistic distribution.

Discussion

This study reveals a substantial disparity in the methods used to register and report postoperative morbidity in leading general surgery journals. Despite the availability of validated tools such as the CDC, their application is often inconsistent and subject to varied interpretations. In the analysis, 16 different uses of the same classification were identified, reflecting efforts to adapt it to the results, a practice that raises concerns given the standardized structure of the CDC.

Furthermore, within a year’s worth of publications in top surgical journals, there is only a 2.8% probability of finding an article that reports morbidity using all core methodological elements.

The low Bayes factor and posterior probability indicate that high overall adherence to core methodological variables is unlikely in the current literature.

The inconsistencies in morbidity reporting stem from the lack of a standardized methodology and consensus on defining, recording, and analyzing postoperative complications.

While classifications such as the ASA score and the Charlson Comorbidity Index are useful and objective tools, they are often insufficient to capture the full clinical context. Contributing factors to subjectivity include differences in severity thresholds, variability in data collection methods, and the inconsistent application (or complete omission) of morbidity classification systems, often replaced by non-standardized, personalized classification systems. Differences in follow-up duration also affect reporting, even though evidence supports both 30- and 90-day follow-up periods^[5]. Additionally, the tendency to report only major (and more accessible) complications further limits data accuracy and comparability.

The omission of minor complications also has a substantial impact on non-complex major procedures, where major complications are exceptional, while a single minor complication could double procedural costs^[14]. Similarly, information regarding the sources used to substantiate morbidity results is rarely provided to readers.

However, as procedures and pathologies become more complex and are associated with increased morbidity, biases are likely to increase due to the absence of a standardized methodology.

The implications of this disparity are numerous, eventually leading to a substantial loss in reproducibility. These inconsistencies may reflect a limited or heterogeneous approach to morbidity reporting, which overlooks minor complications and underscores the secondary role assigned to morbidity. The absence of bias related to morbidity or declared competing interests indicates a limited awareness of the need for objectivity in reporting morbidity outcomes.

These findings highlight the need for consensus-based, standardized protocols for reporting postoperative morbidity. Prior efforts toward standardization include the CLASSIC trial reporting guidelines^[15], which provide a framework for defining complications, applying classification systems, and establishing time frames in clinical trials; the RECOvER Checklist, developed by the ERAS and ERAS USA societies to align complication reporting with recovery outcomes^[7]; and the ISGPS classifications for pancreatic surgery^[16], which exemplify the growing consensus on the need for reproducible reporting tools in specific surgical contexts.

Nevertheless, future research should aim to develop a comprehensive and unified reporting methodology and to validate classification systems such as the CDC and CCI across surgical disciplines. Evaluating their long-term impact on patient outcomes, cost-efficiency, and resource allocation is essential. The implementation of universal morbidity auditing may also enhance surgical practice by identifying areas for improvement, optimizing care delivery, and supporting evidence-based decision-making.

To enhance morbidity reporting, validated classification systems should be required in outcome studies, with transparency promoted through comprehensive data reporting. Standardizing postoperative morbidity data collection and implementing external audits, alongside fostering a culture of accountability, will help establish complication reporting as a quality improvement tool.

The primary limitation of this study lies in the selection of the top six journals, as this could limit the generalizability of these results. Similarly, one year of results is sufficient for a first overview and preliminary conclusions but opens to more extended studies and impels to a need of an external audit of morbidity results. Additionally, variability in surgical populations and procedures across studies may influence reported morbidity rates, introducing further heterogeneity. Comparing results across different pathologies could provide valuable insights, but this was not feasible due to inconsistencies in reported complications and variations in classification usage. Additionally, we acknowledge that the absence of reported methodological detail does not necessarily reflect an absence of rigorous practice. However, our study focuses on what is explicitly presented in the published article, as this is the only information accessible to readers and reviewers. From this perspective, unreported elements limit transparency, comparability, and reproducibility. This further underscores the need for standardized reporting frameworks in surgical research. The diverse study types also limited consistent analysis by procedure.

Conclusion

Morbidity in General Surgery publications often falls short of the standards applied to other results. Readers are left to draw conclusions based on complications that, in reality, have not been subjected to any formal quality verification.

The findings of our study underscore the need for evidence-based guidelines for morbidity reporting. This article aims to initiate a standardization effort to improve the consistency and reliability of such reporting. Future steps toward this goal include developing a Delphi consensus process, conducting multicenter collaborations to assess the applicability of findings across diverse clinical settings, and conducting longitudinal studies to evaluate long-term outcomes and morbidity trends.

In this context, we believe that enhancing both the readers’ critical thinking and the authors’ methods with a new approach to morbidity is an essential first step, one that strives to match quality and transparency in methodology found in other published results.

Ethical approval

Not required for this study.

Consent

Not applicable to this study (no patients involved).

Sources of funding

No funding received.

Author contributions

L.G.: conceptualization, data curation, investigation, methodology, validation, visualization, writing – original draft, writing – review & editing; R.D.L.P.L.: conceptualization, data curation, methodology, project administration, supervision, validation, writing – original draft, Writing – review & editing; D.A.D.C., R.A.G., W.S., C.R.P., and I.A.G.D.R.: data curation, investigation, methodology, visualization, writing – review & editing; S.L.D.: data curation, formal analysis, methodology, validation, visualization, writing – original draft, writing – review & editing.

Conflicts of interest disclosure

None.

Guarantor

Ludovica Gorini and Roberto De La Plaza Llamas.

Research registration unique identifying number (UIN)

Not applicable to this study (no patients involved).

Provenance and peer review

Not commissioned, externally peer-reviewed.

Presentation

Some data from this study were included in the following communications:

• XXIV Reunión Nacional de Cirugía, Alicante, Spain, 23–26 October 2023 (www.elsevier.es/es-revista-cirugia-espanola-36-congresos-xxiv-reunion-nacional-cirugia-159-sesion-gestion-de-calidad-7690-comunicacion-clasificacion-de-la-morbilidad-posoperatoria-94146; www.elsevier.es/es-revista-cirugia-espanola-36-congresos-xxiv-reunion-nacional-cirugia-159-sesion-gestion-de-calidad-7690-comunicacion-modalidades-de-recogida-de-datos-94148).

• “Metodología para el registro, comunicación y publicación de las complicaciones postoperatorias,” Webinar series, American College of Surgeons, Spain Chapter, 13 December 2023.

• XXXV Congreso Nacional de Cirugía, Madrid, Spain, 4–7 November 2024 (https://www.elsevier.es/es-revista-cirugia-espanola-36-congresos-xxxv-congreso-nacional-cirugia-170-sesion-gestion-de-calidad-8329-comunicacion-aplicabilidad-de-un-modelo-basico-102626#).