What is the relationship between hospital caseload and outcome in incisional hernia repair?—A registry-based analysis of 55,584 patients

Abstract

Introduction

There is an ongoing controversal debate about whether the hospital volume (hospital case load) or the surgeon volume has a greater influence on outcome. The implications of high surgeon volume for the outcome have been demonstrated for ventral and incisional hernia repair. This analysis of data from the Herniamed Registry now aims to assess the relationship between hospital volume/hospital case load and outcome in incisional hernia repair.

Methods

To calculate the caseload all repairs carried out in the centers and which met the inclusion criteria were included. The annualized number (based on the time difference between the first and last entry for the respective center) of repairs per center was used for the following categories: < = 20 procedures per year, > 20 - < = 40 procedures per year, > 40 procedures per year. The association of confirmatory defined patient- and procedure-related characteristics to the outcome parameters (general, intraoperative and postoperative surgical complications, complication-related reoperations as well as recurrences, pain at rest, pain on exertion, and chronic pain requiring treatment on 1-year follow-up) was analyzed using logistic regression models.

Results

Following patient selection, 55,584 patients were included in analysis of the relation of the hospital volume as well as of other potential confounders to the outcome parameters.

In the caseload group with ≤ 20 incisional hernias per year, the mean number of surgeons was 6.8, whereas in the caseload group > 20 - ≤ 40 per year the mean number was 12.9 surgeons and in the caseload group > 40 incisional hernias per year, the mean number was 23.7 surgeons. The multivariable analysis of the data from the Herniamed Registry demonstrates that a lower case load is associated with a higher risk of postoperative surgical complications and recurrences, but with a lower risk of pain at rest, pain on exertion and chronic pain requiring treatment.

Conclusion

In summary, this analysis of data from the Herniamed Registry demonstrates that a relatively large number of surgeons are involved in the repair of incisional hernias, regardless of the hospital caseload. High hospital volume comes with a price of more surgeons participating resulting in higher postoperative surgical complication and recurrence risk. Low-volume centers seem to manage less severe cases, whereas high-volume centers appear to act as referral centers, treating more complex cases, which might reflect influence of unobserved confounders.

Supplementary Information

The online version contains supplementary material available at 10.1007/s00423-025-03836-z.

Introduction

In surgical settings, there continues to be controversial debate about whether the hospital volume (hospital caseload) or the surgeon volume has a greater influence on outcome. One systematic review demonstrated for different specialties that a high-surgeon volume and specialization are associated with improved outcome, while high-hospital volume is of limited benefit [1]. The implications of high-surgeon volume for the outcome have also been demonstrated for ventral and incisional hernia repair [2–6].

Data analysis from Statewide Planning and Research Cooperative System in New York State from 2003 to 2009 revealed that for 78,267 ventral hernia repairs the majority of variation in hernia recurrence rates was attributable to surgeon-level variation [4]. Hospital characteristics, e.g. volume, academic affiliation, location, were not associated with recurrences [5].

Analysis of data from the National Inpatient Sample in 2014 based on 54,075 patients from 2,049 hospitals showed that elective, open ventral/incisional hernia repairs in high-volume hospitals appeared to be associated with fewer major complications [7]. The authors concluded that hospitals performing a greater number of ventral hernia repairs, despite caring for a more complex patient population, may be associated with better patient outcomes than lower volume hospitals [7].

A European Hernia Society working group has not specified any minimum caseload for accreditation and certification of hernia centers [8]. However, the German Hernia Society and the German Society of General and Visceral Surgery have specified a minimum caseload of 50 incisional hernia repairs per year for a certified hernia center [9]. The Italian Society of Hernia and Abdominal Wall Surgery has specified for certified hernia centers a minimum caseload of 25 open and 25 laparoscopic incisional hernia repairs [10].

This analysis of data from the Herniamed Registry now aims to identify the potential association of the hospital volume/hospital caseload to the outcome in incisional hernia repair.

Materials and methods

Herniamed is an internet-based hernia registry into which 836 hospitals and independent surgeons in private practice in Germany, Austria and Switzerland have entered prospective data on their routine hernia repairs (from 1 September 2009 until 31 January 2022). All participating hospitals and independent surgeons in private practice sign a cooperation contract in which they commit to entering data completely and correctly into the Herniamed database on all hernia repairs. As a prerequisite for participation in the hernia registry, patients must sign a special consent form, consenting to documentation of their data and to follow-up by the treating hospital and/or surgeon. The patient must not be documented in the Herniamed Registry without this special consent. As part of the information provided to patients regarding their participation in the Herniamed Registry, they are told to inform the treating hospital or surgeon about any problems or complications that occur. If problems or complications occur, the patient should contact the treating hospital or treating surgeon to arrange for clinical examination.

All intraoperative (bleeding, organ injuries, bowel injuries, bladder injuries, nerve injuries, others) and postoperative surgical complications (bleeding, seroma, prolonged ileus, bowel injuries, wound healing disorders, infection) as well as complication-related reoperations are recorded for up to 30 days after surgery. The patient is asked once again about any postoperative complications when they and their general practitioner are sent a questionnaire on 1-year follow-up. On 1-year follow-up the general practitioner and patient are asked about any recurrence or bulging, pain at rest, pain on exertion or chronic pain requiring treatment. The pain intensity is measured with the numeric rating scale (NRS). Clinical examination should be undertaken to investigate any problems reported by the patient.

Based on the inclusion criteria, patients were selected for analysis from the processed data available for a total of 973,469 cases (Fig. 1). The inclusion criteria were as follows:

• incisional hernias

• The patient master data as well as procedure-related data have the entry-state key “complete”, i.e. only completely documented cases regarding the mandatory entries.

• Only laparoscopic and open procedures

• Hospitals and surgeons documented in the Herniamed Registry for at least one year.

• Patients with minimum age of 16 years (only valid age entries)

• Only approved meshes

• Operation date up to, and including, December 31, 2020

• Patients with 1-year follow-up

Fig. 1.

Flowchart of patient inclusion

All analyses were performed with the software SAS 9.4 (SAS Institute Inc. Cary, NC, USA) and were intentionally calculated to a full significance level of 5%, i.e. they were not corrected in respect of multiple tests.

Single outcome variables and potential confounders (risk factors, complications) were summarized as global variables. A general, intra- or postoperative complication or risk factor was deemed present if at least one single item applied.

To calculate the caseload all repairs carried out in the centers which met the first three inclusion criteria were included. The annualized number (based on the time difference between the first and last entry for the respective center) of repairs per center was used for the following categories:

•  < = 20 procedures per year

•  > 20 - < = 40 procedures per year

•  > 40 procedures per year

In univariate analyses, all categorical patient data are reported as absolute and relative frequencies by caseload category. For continuous data, the mean and standard deviation and in some cases minimum, maximum, median as well as the 25% and 75% quartiles (Q1, Q3) are given. Chi-square test and ANOVA were performed for categorical and continuous parameters, respectively.

Main focus wase on logistic regression models taking confirmatory defined patient- and procedure-related characteristics into account as potential confounders on the outcome parameters (general, intraoperative and postoperative complications, complication-related reoperations, recurrences as well as pain at rest, pain on exertion, and chronic pain requiring treatment after one year). All pairwise odds ratios are given with corresponding 95% confidence intervals. The 10-year odds ratio is given for the continuous variable age and a 5-point odds ratio for BMI.

Apart from the caseload categories [< = 20/year/> 20 - < = 40/year/> 40/year], other potential influencing parameters included in multivariable analyses were:

• Age in years

• BMI in kg/m²

• Gender ([male/female]

• ASA [I/II/III-IV]

• Defect size [W1 (< 4 cm)/W2 (> = 4–10 cm)/W3 (> 10 cm)]

• Surgical technique [laparoscopic/open]

• EHS classification [medial/lateral/combined]

• Preoperative pain [yes/no/unknown]

• Drains [yes/no]

• Urgency [elective/emergency]

• Recurrence [yes/no]

• Presence of risk factors [yes/no]

as well as postoperative complications on analysis of pain on follow-up.

Risk factors applied if at least one of the following risk factors was present:

• COPD

• Diabetes mellitus

• Aortic aneurysm

• Immunosuppression

• Corticoids

• Smoking

• Coagulopathy

• Antithrombotic medication (discontinued less than seven days previously)

• Coumarin derivatives (Quick/INR not in normal range)

Results

Following patient selection, 55,584 patients (Fig. 1) were included in analysis of the relation of the hospital volume as well as of other potential confounders to the outcome parameters. Of these patients, 10,106 underwent incisional hernia repair in a center with less than 20 procedures (18.2%), 21,854 patients in a center with 20–40 procedures per year (39.3%) and 23,624 patients in a center with more than 40 procedures per year (42.5%).

The relationship between the incisional hernia caseload and number of surgeons involved per site was calculated (Fig. 2). This showed that in the caseload group with ≤ 20 incisional hernias per year, the mean number of surgeons was 6.8 (± 5.3), whereas in the caseload group > 20 - ≤ 40 per year the mean number was 12.9 (± 6.9) surgeons and in the caseload group > 40 incisional hernias per year, the mean number was 23.7 (± 13.2) surgeons.

Fig. 2.

Caseload and number of surgeons involved per site

Unadjusted analysis

Univariate analysis was performed only to show unadjusted differences in the analysis population. Table 1 presents the descriptive statistics as well as the analysis results for the continuous variables age and BMI, while Table 2 shows the corresponding values for the categorical variables. Unadjusted analysis of the relationship between the hospital-volume categories and the patient- and procedure-related variables revealed that there were highly significant differences in virtually all variable scores between the hospital-volume categories. With such a large number of cases even very small differences (between the groups) can be significant, but their clinical relevance is questionable. Table 3 presents the unadjusted analysis of the relationship between the caseload categories and the intraoperative, general and postoperative surgical complications, complication-related reoperations as well as recurrences, pain at rest, on exertion and requiring treatment on 1-year follow-up. Only for recurrences and pain rates on 1-year follow-up was a significant difference found with regard to the caseload categories. A disadvantage was identified in the recurrence rate for patients from a hospital with a caseload of less than 20 procedures. On the other hand, the pain rates were higher for patients from hospitals with a caseload of more than 20 procedures.

Table 1.

Descriptive statistics and unadjusted analysis results for homogeneity between caseload and age and BMI

		Caseload
		< = 20/year	> 20 - < = 40/year	> 40/year
					p
Age [years]	N/Mean ± SD	10106/62.7 ± 13.6	21854/62.8 ± 13.6	23624/62.5 ± 13.6	0.055
BMI [kg/m2]	N/Mean ± SD	10059/29.5 ± 5.9	21776/29.8 ± 6.1	23525/29.5 ± 6.0	<.001

Table 2.

Descriptive statistics and unadjusted analysis results for homogeneity between caseload and categorical confounding variables

		Caseload
		< = 20/year		> 20 - < = 40/year		> 40/year
		n	%	n	%	n	%	p
Gender	Male	5089	50.4	10763	49.2	11962	50.6	0.010
	Female	5017	49.6	11091	50.8	11662	49.4
ASA	I	1312	13.0	2318	10.6	2305	9.8	<.001
	II	5458	54.0	12254	56.1	13055	55.3
	III/IV	3336	33.0	7282	33.3	8264	35.0
Type of access	Laparoscopic surgery	2532	25.1	5584	25.6	5667	24.0	<.001
	Open surgery	7574	74.9	16270	74.4	17957	76.0
Defect size	I (< 4 cm)	4266	42.2	8209	37.6	8418	35.6	<.001
	II (4–10 cm)	4525	44.8	9871	45.2	10971	46.4
	III (> 10 cm)	1315	13.0	3774	17.3	4235	17.9
EHS classification	Medial	7570	74.9	16825	77.0	18342	77.6	<.001
	Lateral	1715	17.0	3388	15.5	3491	14.8
	Combined	821	8.1	1641	7.5	1791	7.6
Preoperative pain	no	3150	31.2	6850	31.3	7379	31.2	<.001
	yes	6377	63.1	13510	61.8	13869	58.7
	unknown	579	5.7	1494	6.8	2376	10.1
Drainage	yes	5439	53.8	12607	57.7	14098	59.7	<.001
	no	4667	46.2	9247	42.3	9526	40.3
Urgency	Elective	9580	94.8	20731	94.9	22555	95.5	0.003
	Emergency	526	5.2	1123	5.1	1069	4.5
Recurrent operation	yes	2053	20.3	4171	19.1	4744	20.1	0.008
	no	8053	79.7	17683	80.9	18880	79.9
Risk factors - total	yes	4155	41.1	9172	42.0	9514	40.3	0.001
	no	5951	58.9	12682	58.0	14110	59.7
- COPD	yes	1017	10.1	2283	10.4	2489	10.5	0.420
	no	9089	89.9	19571	89.6	21135	89.5
- Diabetes	yes	1395	13.8	2899	13.3	3064	13.0	0.116
	no	8711	86.2	18955	86.7	20560	87.0
- Aortic aneurysm	yes	158	1.6	330	1.5	371	1.6	0.861
	no	9948	98.4	21524	98.5	23253	98.4
- Immunosuppression	yes	151	1.5	411	1.9	510	2.2	<.001
	no	9955	98.5	21443	98.1	23114	97.8
- Corticoids	yes	186	1.8	382	1.7	383	1.6	0.314
	no	9920	98.2	21472	98.3	23241	98.4
- Smoking	yes	1247	12.3	2755	12.6	2833	12.0	0.136
	no	8859	87.7	19099	87.4	20791	88.0
- Coagulopathy	yes	217	2.1	511	2.3	536	2.3	0.566
	no	9889	97.9	21343	97.7	23088	97.7
- Antithrombotic medication	yes	1195	11.8	2772	12.7	2748	11.6	0.002
	no	8911	88.2	19082	87.3	20876	88.4
- Anticoagulant medication	yes	340	3.4	724	3.3	691	2.9	0.026
	no	9766	96.6	21130	96.7	22933	97.1

Table 3.

Descriptive statistics and unadjusted analysis results for homogeneity between caseload and outcome variables

		Caseload
		< = 20/year		> 20 - < = 40/year		> 40/year
		n	%	n	%	n	%	p
Intraoperative complications - total	yes	175	1.7	374	1.7	391	1.7	0.844
	no	9931	98.3	21480	98.3	23233	98.3
General complications - total	yes	338	3.3	806	3.7	866	3.7	0.268
	no	9768	96.7	21048	96.3	22758	96.3
Postoperative complications - total	yes	836	8.3	1743	8.0	1865	7.9	0.499
	no	9270	91.7	20111	92.0	21759	92.1
Complication-related reoperations	yes	340	3.4	763	3.5	831	3.5	0.775
	no	9766	96.6	21091	96.5	22793	96.5
Recurrence on 1-year follow-up	yes	584	5.8	1095	5.0	1147	4.9	0.002
	no	9522	94.2	20759	95.0	22477	95.1
Pain on exertion on 1-year follow-up	yes	1688	16.7	3990	18.3	4484	19.0	<.001
	no	8418	83.3	17864	81.7	19140	81.0
Pain at rest on 1-year follow-up	yes	900	8.9	2186	10.0	2582	10.9	<.001
	no	9206	91.1	19668	90.0	21042	89.1
Pain requiring treatment on 1-year follow-up	yes	732	7.2	1754	8.0	2004	8.5	<.001
	no	9374	92.8	20100	92.0	21620	91.5

Multivariable analysis

Intraoperative complications

The results of the model used for analysis of the relation to the patient- and procedure-related variables are presented in Suppl. 1 (model fit: p < 0.001). The risk of intraoperative complications was significantly associated with defect size, drainage, type of access, age (in each case p < 0.001) and recurrent operations (p = 0.006). No significant relationship was found between the caseload categories and the intraoperative complications.

General complications

The analysis results for the general complication rate are shown in Suppl. 2 (model fit: p < 0.001). The general complications were significantly associated with defect size, urgency, ASA, age, presence of at least one risk factor, drainage, EHS classification and type of access (in each case p < 0.001). No significant relationship was found between the caseload categories and the general complications.

Postoperative surgical complications

The analysis results for the postoperative surgical complication rate are presented in Suppl. 3 (model fit: p < 0.001). Onset of postoperative complications was significantly associated with the defect size, type of access, BMI, presence of at least one risk factor, drainage, urgency, EHS classification, ASA, age (in each case p < 0.001), caseload (p = 0.003) and recurrent operations (p = 0.022).

A lower caseload (< = 20/year vs > 40/year: OR = 1.162 [1.065; 1.268], p < 0.001; < = 20/year vs > 20 - < = 40/year: OR = 1.121 [1.026; 1.224], p = 0.011) was related to a higher postoperative surgical complication risk.

With regard to comparison of the caseload categories, these model results correspond to the following differences: on comparing ` < = 20/year` vs. ` > 40/year` (total prevalence 7.99%, OR = 1.162), there would be around 85 postoperative surgical complications for every 1,000 hernia repairs in patients from centers with less than 20 procedures per year versus 74 for every 1,000 repairs in patients from centers with more than 40 procedures per year. On comparing ` < = 20/year` vs ` > 20 - < = 40/year` (total prevalence 8.07%), an OR of 1.121 corresponds to around 85 postoperative surgical complications for every 1,000 hernia repairs in patients from centers with less than 20 procedures per year and 76 for every 1,000 repairs in patients from centers with 20–40 procedures per year.

Complication related reoperation

The analysis results for the complication-related reoperations are shown in Suppl. 4 (model fit: p < 0.001). The risk of reoperation was significantly associated with defect size, type of access, drainage, presence of at least one risk factor, urgency, EHS classification, ASA, and BMI (in each case p < 0.001). No significant relationship was found between the caseload categories and the complication-related reoperation rate.

Recurrence on 1-year follow-up

The multivariable analysis results for recurrence on 1-year follow-up are illustrated in Suppl. 5 (model fit: p < 0.001). The following variables were found to be related to recurrent operations: EHS classification, BMI, ASA, caseload, type of access, defect size (in each case p < 0.001), age (p = 0.005), preoperative pain (p = 0.010), gender (p = 0.019), urgency (p = 0.023) and drainage (p = 0.024).A lower caseload (< = 20/year vs > 40/year: OR = 1.221 [1.101; 1.354], p < 0.001; < = 20/year vs > 20 - < = 40/year: OR = 1.168 [1.053; 1.296], p = 0.003) was associated with a higher risk of recurrence.

With regard to comparison of the caseload categories, these model results correspond to the following differences: on comparing ` < = 20/year` vs. ` > 40/year` (total prevalence 5.14%, OR = 1.221) there would be around 57 recurrences for every 1,000 hernia repairs in patients from centers with less than 20 procedures per year versus 47 for every 1,000 repairs in patients from centers with more than 40 procedures per year. On comparing ` < = 20/year` vs ` > 20 - < = 40/year` (total prevalence 5.26%), an OR of 1.168 corresponds to around 57 recurrences for every 1,000 hernia repairs in patients from centers with less than 20 procedures per year and 49 for every 1,000 repairs in patients from centers with 20–40 procedures per year.

Pain at rest on 1-year follow-up

The analysis results for pain at rest on 1-year follow-up are summarized in Suppl. 6 (model fit: p < 0.001). There was a significant relationship between pain at rest and age, preoperative pain, gender, postoperative complications, EHS classification, urgency, defect size, ASA, caseload, recurrent operations, BMI, and drainage (in each case p < 0.001).

A lower caseload (< = 20/year vs > 40/year: OR = 0.802 [0.740; 0.870], p < 0.001; > 20 - < = 40/year vs > 40/year: OR = 0.909 [0.855; 0.967], p = 0.002; < = 20/year vs > 20 - < = 40/year: OR = 0.882 [0.812; 0.958], p = 0.003) was associated with a lower risk of pain at rest.

With regard to comparison of the caseload categories, these model results correspond to the following differences: on comparing ` < = 20/year` vs. ` > 40/year` (total prevalence 10.32%, OR = 0.802), there would be around 93 cases of pain at rest on follow-up for every 1,000 hernia repairs in patients from centers with less than 20 procedures per year versus 113 cases for every 1,000 repairs in patients from centers with more than 40 procedures per year. On comparing ` > 20 - < = 40/year` vs ` > 40/year` (total prevalence 10.48%), an OR of 0.909 corresponds to around 100 cases with pain at rest on follow-up for every 1,000 hernia repairs in patients from centers with 20–40 procedures per year and 109 for every 1,000 repairs in patients from centers with more than 40 procedures per year. Furthermore, on comparing ` < = 20/year` vs ` > 20 - < = 40/year` (total prevalence 9.64%, OR = 0.882), there would be around 91 cases of pain at rest on follow-up for every 1,000 hernia repairs in patients from centers with less than 20 procedures per year and 102 for every 1,000 repairs in patients from centers with 20–40 procedures per year.

Pain on exertion on 1-year follow-up

The analysis results for pain on exertion on 1-year follow-up are summarized in Suppl. 7 (model fit: p < 0.001). It was significantly associated with age, gender, preoperative pain, EHS classification, postoperative complications, defect size, urgency, recurrent operations, drainage, BMI, type of access, caseload (in each case p < 0.001), presence of at least one risk factor (p = 0.003) and ASA (p = 0.030).

Here too, a lower caseload (< = 20/year vs > 40/year: OR = 0.861 [0.808; 0.917], p < 0.001; < = 20/year vs > 20 - < = 40/year: OR = 0.902 [0.846; 0.962], p = 0.002) was associated with a lower risk of pain on exertion.

With regard to comparison of the caseload categories, these model results correspond to the following differences: on comparing ` < = 20/year` vs. ` > 40/year` (total prevalence 18.30%, OR = 0.861), there would be around 172 cases of pain on exertion on follow-up for every 1,000 hernia repairs in patients from centers with less than 20 procedures per year versus 194 cases for every 1,000 repairs in patients from centers with more than 40 procedures per year. On comparing ` > 20 - < = 40/year` vs ` > 40/year` (total prevalence 17.75%), an OR of 0.902 corresponds to around 170 cases with pain on exertion on follow-up for every 1,000 hernia repairs in patients from centers with 20–40 procedures per year and 185 cases for every 1,000 repairs in patients from centers with more than 40 procedures per year.

Chronic pain requiring treatment on 1-year follow-up

The analysis results for chronic pain requiring treatment are presented in Suppl. 8 (model fit: p < 0.001). A significant relation was identified here between age, preoperative pain, gender, postoperative complications, EHS classification, recurrent operations, ASA, defect size, urgency, drainage, presence of at least one risk factor (in each case, p < 0.001), type of access (p = 0.001), caseload (p = 0.002) and BMI (p = 0.004).

Once again, a lower caseload (< = 20/year vs > 40/year: OR = 0.849 [0.776; 0.929], p < 0.001; < = 20/year vs > 20 - < = 40/year: OR = 0.901 [0.823; 0.987], p = 0.025) was associated with a lower risk of pain requiring treatment.

With regard to comparison of the caseload categories, these model results correspond to the following differences: on comparing ` < = 20/year` vs. ` > 40/year` (total prevalence 8.12%, OR = 0.849), there would be around 75 cases of pain requiring treatment on follow-up for every 1,000 hernia repairs in patients from centers with less than 20 procedures per year versus 87 cases for every 1,000 repairs in patients from centers with more than 40 procedures per year. On comparing ` > 20 - < = 40/year` vs ` > 40/year` (total prevalence 7.78%), an OR of 0.901 corresponds to around 74 cases with pain requiring treatment on follow-up for every 1,000 hernia repairs in patients from centers with less than 20 procedures per year and 81 cases for every 1,000 repairs in patients from centers with 20–40 procedures per year.

Discussion

This manuscript presents a retrospective, registry-based analysis investigating the association between hospital case load and surgical outcomes in incisional hernia repair. The authors assess whether high-volume centers achieve superior clinical outcomes compared to lower-volume institutions. The study addresses an important aspect of surgical quality and healthcare policy, providing evidence that may support centralization strategies for complex abdominal wall procedures. To define the three volume thresholds was difficult. The German certification system for hernia centers requires 30 incisional hernia operations per year. In the literature two studies conclude that a minimum of 20–25 procedures per year, per surgeon are related to lower reoperations rates and lower perioperative costs [10].

This multivariable analysis of data from the Herniamed Registry demonstrates, while taking account of the potential confounding exerted by other patient- and procedure-related characteristics, that the caseload categories have a significant relation to the postoperative surgical complications, recurrences, pain at rest, pain on exertion, and to chronic pain requiring treatment. A lower caseload is associated with a higher risk of postoperative surgical complications and recurrences, but with a lower risk of pain at rest, pain on exertion, and chronic pain requiring treatment.

An unfavorable association is found between chronic pain requiring treatment and preoperative pain, female gender, postoperative complications, recurrence repair, higher ASA score, larger defects, the use of drains, and the presence of risk factors.

A favorable relationship is seen between chronic pain requiring treatment and higher age, medial EHS classification, emergency procedures, open surgical techniques, higher BMI, and a lower caseload.

Within the individual caseload categories, the repairs were performed on average by a varying number of surgeons (Fig. 2). For the caseload category ‘ ≤ 20 per year’, the average number of surgeons was 6.8, for ‘ > 20 - ≤ 40 per year’ it was 12.9 and for ‘ > 40 per year’ it was 23.7. For the given caseload, the number of incisional hernias repaired per surgeon per year was on average between 2.9 and 1.7. As such, with regard to the number of repairs carried out, no high-volume effect for the surgeon is identifiable, leading to the positive effect on outcome as reported in the literature [2–6]. For certification of hernia centers by surgical societies, in addition to the requirement for the total caseload per center per year (hospital volume/hospital caseload), there should also be a requirement for the minimum number of operations performed per surgeon (surgeon volume). The certification regulations of the German Hernia Society and the German Society for General and Visceral Surgery stipulate that up to four designated surgeons should be present for all hernia repairs [9]. In this way, it is possible to ensure that both a high-hospital volume and high-surgeon volume are assured in hernia centers.

It is not easy to interpret the findings of the analysis presented here. For the caseload category ≤ 20 incisional hernia repairs, one would have expected in addition to higher postoperative surgical complication rates and recurrence rates also higher rates of pain at rest, pain on exertion, and chronic pain requiring treatment. Unadjusted comparison of the patient- and diagnosis-related characteristics (Table 2) shows for the caseload category ≤ 20 a proportion of 13% of defect sizes > 10 cm and for the caseload category > 40 per year it shows a proportion of 17.9% (p < 0.001).

The proportion of open surgical techniques was 74.9% for the caseload ≤ 20 per year versus 76.0% for the caseload > 40 per year (p < 0.001). Hence, in hospitals with a caseload ≤ 20 incisional hernia repairs significantly fewer open repairs of incisional hernias with defect sizes > 10 cm were performed [11]. Both these factors may result in a lower rate of pain at rest, on exertion, and chronic pain requiring treatment on 1-year follow-up.

There are no better studies available for incisional hernias in terms of hospital caseload. However, the present results clearly show, that it is not so much the number of cases in a clinic that matters, but rather the experience of the individual surgeons. An improvement in the results of incisional hernia surgery is best achieved by surgeons with extensive experience [12].

Registry analyses have limitations. Missing or incorrect data can negatively impact the findings. All responsible surgeons participating in Herniamed sign a cooperation contract where they commit to entering data completely and correctly on patients undergoing hernia repair. Missing data are highlighted by the software. On 1-year follow-up the patient and their general practitioner are sent a questionnaire asking once again about any perioperative complications. At the time of certification of hernia centers, the auditors can check for complete and correct data entry.

Missing 1-year follow-up data for a relevant proportion of patients is another limitation. However, calculation of standardized differences between patients with and without follow-up does not point to any selection bias (Fig. 3). With such a large number of patients small differences between groups can be significant, but the clinical relevance is questionable. Moreover, there is a major difference in the size of the groups.

Fig. 3.

Scatter plot of standardized differences of patients with and without follow-up

In summary, this analysis of data from the Herniamed Registry demonstrates that a relatively large number of surgeons are involved in the repair of incisional hernias, regardless of the hospital caseload. This means that the goal of a high-volume hospital with high surgeon volume cannot be achieved.

Hence, the findings of the Herniamed Registry analysis tend to be more representative of low-surgeon volume settings, posing a higher risk of unfavorable outcomes. For the intraoperative complications, general complications and complication-related reoperations, there is no significant association with the hospital caseload (hospital volume) but there is an unfavorable relationship between a low- hospital caseload and the postoperative surgical complications and recurrences on 1-year follow-up. Low-volume hospitals have a favorable relation to pain at rest, pain on exertion and chronic pain requiring treatment.

Low-volume centers seem to manage less severe cases, whereas high-volume centers appear to act as referral centers, treating more complex cases. More studies are needed to show that a high-volume center per se does not produce better results in incisional hernia surgery, but only when the surgeons involved are high-volume surgeons.

Supplementary Information

Below is the link to the electronic supplementary material.

Author contributions

F.K wrote the main manuscript text and D.A prepared Figs. 1, 2 and 3, Tables 1, 2 and 3 and all supplements. All authors reviewed the manuscript.

Data availability

The data can obtained upon request from the corresponding author.

Declarations

Ethical approval

Only cases of routine hernia surgery were documented in the Herniamed Registry and all patients have signed a special informed consent declaration agreeing to participate.

The Herniamed Registry has ethical approval (BASEC No. 2016 - 00.123, 287/2017BO2).

Human and animal rights

This article does not contain any study with animals performed by any of the authors.

Informed consent

All patients with routine hernia surgery documented in Herniamed Registry have signed an informed consent declaration agreeing to participate.

Competing interests

The authors declare no competing interests.