What is the current provision of service for gastrostomy insertion in England?

Abstract

Background

Significant morbidity and mortality can be associated with gastrostomy insertion, likely influenced by patient selection, indication and aftercare. We aimed to establish what current variation in practice exists and how this has improved by comparison to our previously published British Society of Gastroenterology survey of 2010.

Methods

We approached all National Health Service (NHS) hospitals in England (n=198). Email and web-based questionnaires were circulated. These data were correlated with the National Endoscopy Database (NED).

Results

The response rate was 69% (n=136/198). Estimated Percutaneous Endoscopic Gastrostomy (PEG) placements in the UK are currently 6500 vs 17 000 in 2010 (p<0.01). There is a dedicated PEG consultant involved in 59% of the centres versus 30% in 2010 (p<0.001). Multidisciplinary team meeting (MDT) discussion occurs in 66% versus 40% in 2010 (p<0.05). Formal aftercare provision occurs in 83% versus 64% in 2010 (p<0.001). 74/107 respondents (69%) reported feeling pressurised to authorise a gastrostomy.

Conclusion

This national survey, validated by the results from NED, demonstrates a reduction of over 60% for PEG insertion rates compared with previous estimates. There has also been an increase in consultant involvement, MDT discussion and aftercare provision. However, two-third of responders described ‘pressure’ to insert a gastrostomy. Perhaps further efforts are needed to include and educate other specialty teams, patients and next of kin.

WHAT IS ALREADY KNOWN ON THIS TOPIC

• Among endoscopic procedures, gastrostomy insertion is reported to have a high morbidity and mortality. The National Confidential Enquiry into Patient Outcome and Death in 2004 reported that approximately 17 000 PEGs were undertaken annually in the UK. In those who had died within 30 days following the procedure, 19% were deemed futile.

WHAT THIS STUDY ADDS

• In the last 10 years, there has been a significant reduction in the number of PEG tubes inserted. There has also been an increase in consultant involvement, MDT discussions and aftercare provision.

HOW THIS STUDY MIGHT AFFECT RESEARCH, PRACTICE OR POLICY

• This study suggests gastrostomy protocols should be incorporated into all hospitals with a gastrostomy service and all patients should be discussed in an MDT. This may help reduce the number of unnecessary procedures and provide better patient care.

Introduction

A gastrostomy provides a direct feeding route into the stomach. It is typically used for patients unable to have sufficient intake orally for a prolonged period. Percutaneous Endoscopic Gastrostomy (PEG) is the most common method in the UK used for gastrostomy insertion. Annual PEG incidence in the UK has previously been reported at approximately 17 000.^{1 2}

The National Confidential Enquiry into Patient Outcome and Death in 2004 found that Patients having PEG insertions had the highest 30-day mortality rate of all endoscopic procedures at 6%. Retrospective case note review suggested that 19% should not have had the procedure, as it was considered futile.¹ The decision to proceed with a gastrostomy varies greatly among health professionals. It is influenced not only by clinical judgement but also by individual beliefs of the clinician, other allied healthcare professionals, the patient and family/carers.^{3 4} Large, good-quality studies are scarce and there is not always sufficient evidence to suggest it will be of benefit, particularly in those with dementia or multiple comorbidities.⁵ Large-scale information on how gastrostomy procedures are performed across trusts is needed for quality assurance and to see how practice varies. This information could potentially help unify best practice and influence guidelines on whom and how the procedure is performed.

We previously published a British Society of Gastroenterology (BSG) endorsed national survey in 2010.² We repeated this in 2021 to get an update of how practice has changed. We have also analysed the UK National Endoscopy Database (NED), which began in 2013. NED is a centralised database that currently holds data from 445 JAG registered sites (86%). It automatically obtains set data from each procedure as it is entered into local endoscopy reporting systems.^{6 7} Our aim is to use these two large sets of data to further our understanding of gastrostomy insertions and the process involved in the patient’s care before, during and after.

Method

Survey

An email questionnaire about PEG service provision was sent out to 198 hospitals in England in January 2021. Email addresses were identified through the BSG 2020 consultant directory. An online questionnaire was also produced. All hospitals were contacted and attempts to follow-up were made by email and telephone.

Questions replicated those previously asked in 2010; including the number of procedures performed annually, who inserts them, indications as well as the processes involved in selecting patients and any aftercare provisions. Additional questions were asked to determine how often disagreements occur and whether they ever felt pressurised to authorise the procedure.

National Endoscopy Database

Data were retrospectively analysed in excel. Only those with PEG placements mentioned as a treatment or indication from the period January 2019 to November 2020 were included in the analysis. The information assessed included age and gender of the patient, whether they were an inpatient or outpatient and urgency of the procedure. Also recorded was job role of the endoscopist and the sedation and dosage used.

Statistical tests used were Fishers exact test when comparing data between years with expectant small numbers and χ² test when comparing categorical variables. Results where p<0.05 were classed as significant.

Results

Survey

Of 126 trusts accounting for 198 hospitals in England (72 trusts have one hospital each, 39 trusts=two hospitals, 13 trusts=three hospitals and two trusts=four hospitals). We received responses from 105/126 (83)% of all trusts surveyed, some trusts were offering different gastrostomy services between hospitals. In total, we received 107 responses which covered 136/198 (69%) of hospitals. A total of 61/107 (57%) responses were from gastroenterology consultants, 31/107 (29%) nutrition nurses, 7/107 (6.5%) nurse endoscopists (6.5%), 7/107 (6.5%) dieticians and 1/107 (1%) specialist doctors.

Who inserts PEGs in your hospital?

Of 102/107 (95%) hospitals reported that consultant gastroenterologists were involved in inserting PEGs. Of 34/107 (32%) had surgeons inserting, 29/107 (27%) nurse endoscopists and 29/107 (27%) involved specialist registrars. This compared with in 2011: 100%, 33%, 36% and 51%, respectively.

As in 2010, participants were also given the option of specifying other specialities that participated in PEG insertion. 17/107 (16%) commented that nutrition nurses were also involved, 1/107 (1%) care of the elderly physicians and 2/107 (2%) associate specialists. In 2010, no nutrition nurses were mentioned, care of the elderly physicians 3% and general practitioners<1%.

How many PEGs were inserted in your hospital last year?

Participants gave an estimate of how many PEGs in their hospital were inserted annually. This was compared with NED in 2019 (table 1). In 2010, 4/215 hospitals (2%) reported doing over 150 PEG placements a year, in 2021, it was 7/107 (6.5%) (p<0.05).

Table 1.

Number of PEGs inserted per year in each hospital

Number of PEGs per year	Survey (2021) n=107	NED (2019) n=155
0–25	28 (26%)	102 (66%)
26–50	36 (34%)	34 (22%)
51–75	12 (11%)	7 (5%)
76–100	14 (13%)	7 (5%)
101–125	8 (7%)	2 (1%)
126–150	1 (1%)	0
151–175	1 (1%)	1 (1%)
176–200	4 (4%)	1 (1%)
>200	2 (2%)	1 (1%)
Not answered	1 (1%)

NED, National Endoscopy Database; PEG, Percutaneous Endoscopic Gastrostomy.

Do you have access to radiological-guided gastrostomies?

Of 86/107 (80%) of hospitals have access to radiological-guided gastrostomies (RIGs), this was 78% in 2010. In 2021, PEG insertion was more common than RIGs in 53/107 (61%), PEG was equal to RIG insertion, 16/107 (15%), PEGs were less common than RIGs, 18/107 (17%).

Indications

Of 104/107 (97.2%) of sites inserted gastrostomies for dysphagic strokes, 103/107 (96%) for neurodegenerative disorders, 22/107 (20.6%) for dementia, 96/107 (89.7%) for head injury, 69/107 (64.5%) for oropharyngeal malignancy and 71/107 (66.36%) performed venting PEGs.

PEG provisions

Hospitals were asked if they had access to certain provisions. This was compared with the responses from 2010 (table 2). In addition, 64/107 (60%) had a predetermined PEG assessment form. Local guidelines were available in 47/107 (44%) of sites. Of 7/107 (7%) of responses used prognostic scoring systems in their assessment of PEG placement. Of 6/107 (6%) of these were the Sheffield gastrostomy score,⁸ 1/107 (1%) was not recorded. Of 75/107 (70%) had a set policy for patients who cannot consent.

Table 2.

PEG provisions in hospital 2010 and 2021

Provision	2010 n=216		2021 n=107		P value
PEG nurse	99	46%	60	56%	0.1
PEG consultant	65	30%	63	59%	<0.001
Dedicated PEG MDT	129	60%	47	44%	0.01
Patient discussion in MDT	86	40%	71	66%	<0.05
A dedicated PEG endoscopy list	52	24%	36	34%	0.08
A dedicated aftercare service for PEG patients	138	64%	89	83%	<0.001
PEG referral form	112	52%	67	63%	0.08

MDT, multidisciplinary team meeting; PEG, Percutaneous Endoscopic Gastrostomy.

How are patients assessed?

Centres were asked which healthcare professionals were routinely involved in assessing a patient prior to their PEG placement. All patients were assessed by at least one professional, prior to insertion. The assessment included an endoscopy nurse assessment in 12/107 (11%), a nutrition nurse review in 68/107 (64%), a gastroenterology doctor review (either consultant or registrar) in 61/107 (57%), a dietician review 75/107 (70%) and a speech and language specialist review 71/107 (66%).

Prophylactic antibiotic use

Of 102/107 (95%) reported using antibiotics either pre, during or immediately post-PEG insertion. The most common antibiotic used was intravenous coamoxiclav with 51/102 (50%) using as sole prophylaxis. There was one report of using cotrimoxazole via the PEG.

What is the time delay between PEG referral and insertion?

In 2021: 27/107 (25%) were requested immediately versus 14% in 2010 (p value: 0.02). Of 43/107 (40%) were dependent on the indication in 2021 compared with 33% in 2010 (p value: 0.27).

Disagreement and pressure

Of 70/107 (65%) reported a referral was refused 0–20% of the time, 21/107 (20%) refused 21–40%, 7/107 (7%) refused 41–60% and 1/107 (1%) refused a referral 61–80% of the time. Reasons for refusal included comorbidity, perceived futility, family/carers wishes and patient’s wishes. Only 11/107 (10%) of respondents reported never experiencing a disagreement regarding the decision to insert a PEG. The large majority 91/107 (85%) reported this only occurred between 1% and 25%, one response reported it happening between 26% and 50% of the time.

Of 74/107 (69%) of responders reported having felt pressurised to authorise a PEG. Of 55/107 (77.5%) had felt pressurised by other specialist teams, 9/107 (13%) by their own team, 53/107 (75%) by family members and 23/107 (32%) by the patient. Feeling pressurised was not statistically related to job role, p=0.172.

National Endoscopy Database

Demographics

Of 7758 PEG placements were performed from January 2019 to November 2020. This was from 127 trusts, which included 196 hospitals. In 2019, England represented 94% of all data, Scotland: 4% and Wales 2% (table 3).

Table 3.

Total number of PEGs on NED performed in 2019

	Number of PEGs	Per site	Per trust
England (total)	4105	155	111
Scotland	203	12	5
Wales	85	7	2
Total	4393	174	118
Median no of PEGs performed		15 (range: 1–210)	25 (range: 1–210)

NED, National Endoscopy Database; PEG, Percutaneous Endoscopic Gastrostomy.

Between 1 and 150 PEGs were performed by each endoscopist in 2019 with a median average of 2 per endoscopist. Only 93/957 (10%) of endoscopists did more than 10 procedures a year.

In England in 2019, 155 sites for 111 trusts inputted their data on PEG insertion into NED. They recorded 4105 PEG insertions. This closely correlates to the data collected in the gastrostomy survey 2021 (136 hospitals from 105 trusts, estimating between 4073 and 6925 procedures).

In 2002–2003, it was estimated approximately 17 000 PEGs were performed annually in the UK (total population 59 636 700).⁹ This estimate was taken from data collected from England only; however, these data represented 94% of all the UK.¹ This is the same representation in NED. In 2010, the national survey also estimated 17 000 PEGs performed annually (total population 62 759 500).⁹ NED recorded 4393 procedures performed in the UK in 2019. This was from 118/148 trusts and 174/254 sites (69%). This gives an estimation of 6 500 PEGs performed in the UK annually (total population 67 081 000).⁹ Which is >60% reduction from 2010 (p <0.001).

Of 4367/7758 (56%) of PEG placements were in men, 3312/7758 (47%) women, 82/7758 (1%) unclassified. The median average age of the patient was 64 (range: 0–102). Gastroenterologists performed the procedure in 6698/7758 (86%), nurse endoscopists in 302/7758 (4%), gastrointestinal surgeons in 417/7758 (5%) and one procedure done by a general practitioner, 339/7758, (5%) were not specified. Of 4679/7758 (60%) of PEGs were done as an inpatient, 2298/7758 (30%) recorded as outpatients, 781/7758 (10%) not specified. Of 3425/7758 (44%) of cases, the endoscopist documented the procedure as routine, 3704/7758 (48%) urgent, 190/7758 (2%) emergency, 6% not specified.

Sedation

There was no variation between discomfort scores and sedation given. Twelve patients had Entonox but not used in isolation. Midazolam alone was used in 2537/7758 (33%) of cases. The median dose was 2 mg (range: 0.5–9 mg). Midazolam combined with fentanyl was used in 3743 (48%) cases. Median dose of midazolam 2 mg (range: 0.5–10 mg), median fentanyl dose 50 μg (range: 10–200 μg). Midazolam combined with pethidine: 414 (5%) median dose of midazolam 2.5 mg (range: 0.5–10 mg), median pethidine dose 25 mg (range: 10–100 mg). General anaesthetic (GA): 282/7758 (4%). Propofol excluding GA : 80/7758 (1%). No drugs given: 648 (8%). Fentanyl alone: 53 (0.7%). Median dose: 50 μg (range: 2–100 μg). Sedation practice for PEG insertion is summarised in figure 1.

Figure 1.

Pie chart demonstrating sedation practices in PEG insertion. PEG, Percutaneous Endoscopic Gastrostomy.

Discussion

The key finding from the national survey and NED was a significant reduction in PEGs performed over the last 10 years. Based on this new information, there are approximately 6 500 PEGs performed annually in the UK compared with 17 000 when calculated in 2003 and 2010 (>60% reduction).^{1 2} There is also a significant increase in the number of PEG consultants (30%–59% p<0.001) and discussions in aMDT meeting prior to the procedure (40%–66% p<0.05), although a reduction in dedicated PEG MDTs. This suggests there is now a more comprehensive, co-ordinated selection process when deciding who should proceed to a gastrostomy and this may have contributed to a reduction in the number of unnecessary procedures. Previous studies have demonstrated how MDT discussion can reduce 30-day gastrostomy mortality,¹⁰ as well as PEG complications.¹¹ Alternative methods of gastrostomy may account for some of the reduction in PEGs; however, PEGs are still more common than RIGs in the majority of hospitals and compared with 2010, there has been no change in hospital access to RIG. This cannot, therefore, account for the dramatic reduction in procedures. Future studies would be helpful to assess how the initial assessment process varies, depending on gastrostomy insertion technique. The reduction in gastrostomy numbers may partially reflect advancement in treatments that delay or prevent the progression of illness. For example: urgent thrombolysis for those suffering an ischaemic stroke has been shown to have significantly reduced the number of deaths and people left dependent,¹² likely resulting in a reduction of patients who require long-term enteral feeding.

The survey showed an increase in formal gastrostomy aftercare provision (64%–83% p<0.001), indicating more support for patients’ postprocedure, which hopefully will reduce the associated morbidity. There was no significant change in the timing of a gastrostomy after receiving an initial referral, from 2010 to 2021. However, more centres now undertake immediate insertions, and this may reflect the improvement in the selection process, this is also demonstrated by more hospitals having a delay that is dependent on the individual case.

Of 21% of hospitals would still insert a gastrostomy for dementia, this is less than in 2010 (36%). There was, however, often an additional comment explaining this was usually an exception. This suggests that most hospitals are aware of the national guidelines, to generally avoid gastrostomies in this group, due to lack of proven benefit.¹³

Those answering the survey were asked whether they had ever felt pressurised to insert a gastrostomy. Sixty-nine per cent of those answering reported that they had. There was no significant variation between job role and feeling pressure. Pressure was experienced mostly from other specialist teams and family members. This highlights the importance of an MDT, so that the responsibility for a decision is not put on one individual and a holistic approach to the patient can be used.

NED showed 66% of hospitals inserting a small number of PEGs a year (1–25) compared with the estimates given in the survey (26%). It may be that when only small numbers are performed, there was an increased likelihood of overestimating or that centres with smaller numbers were less likely to return the questionnaire and/or have formal records.

NED showed a considerable variation among the number of PEGs performed annually by endoscopist as well as per site and trust. The median number of PEGs performed by an endoscopist yearly is 2 (range: 1–150), with only 10% of endoscopists doing more than 10 procedures a year. There are no studies that have demonstrated an association between the number of PEGs performed and outcomes. For other endoscopic procedures, there are a recommended minimum number of procedures a year to maintain competency.¹⁴ There may be implications to service provision if a minimum requirement is extended to PEG insertion.

Sedation type and amount used when inserting a PEG varied largely, although did not appear to affect discomfort scores. This may reflect the different indications, age and comorbidities of the patient as well as the experience and availability of drugs to the individual endoscopist.

Limitations

The survey is based on an individual’s response, the number of questions is restricted and sometimes subjective, with limited scope to expand, explaining answers. The results only include England; however, the previous well-cited estimate of 17 000 annual UK PEG placements was calculated on only English data, which represented 94% of trusts in the UK.¹ The NED database had some potentially relevant information in free text making analysis difficult. Data were collected during COVID-19; however, participants answered the survey based on usual hospital practice, pre-COVID-19. Data from this were directly compared with the last year of complete NED data (2019) and these closely correlated.

Conclusion

Advances appear to have been made in gastrostomy provision among hospitals in England, including senior input and MDT discussion. There has been a reduction in the number of PEGs inserted in the UK, with an estimate of 6500 annually and a reduction of more than 60% from previous estimates. It may be that the selection process has improved; therefore, fewer unnecessary procedures are being performed. A recommended standard of care for hospitals should be that all patients are discussed in an MDT prior to gastrostomy insertion.

Further work is needed as disagreement still occurs, even experienced staff feel occasional pressure. There appears to have been an increase in knowledge within gastroenterology departments on the potential morbidity and mortality associated with gastrostomy; however, further education on this for other healthcare professionals may be needed and should be incorporated into medical, nursing and allied health professionals’ training days.

Data availability statement

All data relevant to the study are included in the article or uploaded as supplementary information.

Ethics statements

Patient consent for publication

Not applicable.

Ethics approval

Not applicable.