Death Certification Test Site Evaluation

2 Quantitative and Qualitative Evidence from the Test Sites

Introduction

2.1 This chapter provides the quantitative evidence from the test sites under the following headings:

• test site activity, covering information about the numbers of reviews undertaken;
• outcome of reviews; and
• duration of reviews.

2.2 The chapter then provides qualitative findings related to the processes involved within each test site.

Test site activity

The number of reviews completed

2.3 Table 2.1 below sets out statistics regarding the numbers of reviews in both test sites. The source for all statistics shown in this chapter is NHS ISD and the data was extracted from the ISD database on 9 April 2013.

Table 2.1: Number of reviews completed by level and test site

Type of Review	Dumfries & Galloway	Dundee	Total
Level 1 (Basic)	618	505	1123
Level 2 (Comprehensive)	100	65	165
Upgraded from Level 1 to Level 2	9	4	13
Total	727	574	1301

2.4 The total cases reviewed in each area tallies closely with the original estimated targets of 525 level 1s and 70 level 2s and the overall total of 1301 exceeds the original estimate of 1190 reviews.

2.5 Registrars followed a random sampling process to refer a percentage of deaths (excluding Procurator Fiscal referrals) to MRs for review. Registrars were given a list of numbers and instructed to refer any deaths for which the entry number in their register matched or ended in one of those numbers.

2.6 Table 2.1 shows that the MRs in the two test site areas completed 1301 reviews in total, comprising 1123 Level 1 reviews (86% of all reviews), 165 Level 2 reviews (13%) and an additional 13 reviews upgraded from Level 1 to Level 2 (1%).

2.7 There have been more reviews in Dumfries & Galloway than Dundee. The ISD database recorded that the MR in Dumfries & Galloway completed 727 reviews^[4] (618 level 1, 100 level 2 and nine upgraded). In Dundee, the MR completed 574 reviews (505 level 1, 65 level 2 and four upgraded). This shows that the number of cases reviewed in Dundee (574) was a little short of the target for the test site (595).

2.8 Despite the discrepancy between the two test sites in terms of the volume of completed reviews, we should note that the number of reviews that MRs completed was dictated by the number of cases that registrars referred to them. This was in turn dictated by a random sampling method which prompted registrars to refer a percentage of deaths for review. The fact that the MR in Dumfries & Galloway completed more reviews than in Dundee could potentially be due to a range of reasons including differences in the ways the referral processes worked in practice in the two areas or differences in the overall death rate in the two areas.

Table 2.2: Number of reviews completed per month by level and test site

Month of Referral	Dumfries & Galloway			Dundee			Total
	Level 1	Level 2	Upgraded from Level 1 to Level 2	Level 1	Level 2	Upgraded from Level 1 to Level 2
May 2012*	-	-	-	30	3	-	33
June 2012	48	6	-	52	4	-	110
July 2012	97	15	2	34	4	1	153
August 2012	84	9	-	58	5	-	156
September 2012	73	12	2	49	8	1	145
October 2012	106	11	2	44	10	-	173
November 2012	95	7	-	69	4	-	175
December 2012	86	12	-	59	11	-	168
January 2013	29	22	2	76	13	2	144
February 2013	-	6	1	34	3	-	44
Total	618	100	9	505	65	4	1301

*Note: The Dumfries & Galloway test site began in June.

Figure 2.1: Number of reviews completed per month

2.9 Table 2.2 and Figure 2.1 show that there were some fluctuations in the number of reviews completed per month. The median number of reviews completed per month is 149 but the number of reviews completed per month increased as the test sites became more established. In June, the first month when both test sites were operational, 110 reviews were completed but this increased to 153 in July and continued to increase, with the exception of a slight decrease in September, to 173 in October, a peak of 175 in November, and 168 in December. It is likely that the number of reviews peaked in winter months because the death rate is higher at that time of year, and hence the MRs received more referrals in these months. The number of reviews completed per month began to decrease again towards the end of the test sites: in January, 144 reviews were completed^[5] and 44 were completed in February before the end of the test sites on the 14^th of that month.

Outcome of reviews

Figure 2.2: Outcome of reviews

Note: Not available refers to six cases in Dumfries & Galloway which had been referred to the Procurator Fiscal before referral to the MR.

2.10 Figure 2.2 and Table 2.3 show that the majority of MCCDs reviewed (1254 or 96%) were found to be 'in order'. Only 41 MCCDs (3% of all reviews) were found to be not in order. This indicates that within the high level scrutiny of test site conditions the majority of MCCDs had been completed accurately by doctors and is a very positive result.

2.11 We discuss in more detail below the reasons why an MCCD was rated as not in order.

Table 2.3: Outcome of reviews by area

Outcome of Review	Dumfries & Galloway			Dundee			Total
	Level 1	Level 2	Upgraded from Level 1 to Level 2	Level 1	Level 2	Upgraded from Level 1 to Level 2
MCCD is 'In Order'	609	98	9	473	63	2	1254
MCCD is 'Not In Order' and Certifying Doctor issued replacement MCCD	3	2	-	32	2	2	41
Not available*	6	-	-	-	-	-	6
Total	618	100	9	505	65	4	1301

*Note: The six Dumfries & Galloway cases which are not available had previously been referred to the Procurator Fiscal before referral to the MR.

2.12 Table 2.3 indicates that more MCCDs were not in order and required a replacement MCCD in Dundee than Dumfries & Galloway. Thirty-six MCCDs in Dundee (6% of all reviews in that area) were found to be not in order compared with five in Dumfries & Galloway (less than 1% of reviews in that area). There could be a number of reasons for this including different approaches by certifying doctors in each area or a different approach being taken between the MRs in the two areas and we discuss this issue further in Chapter 3 as it highlights the need to ensure consistency among MRs when the process is implemented nationally.

2.13 Looking at all level 1 and level 2 reviews across both areas, we found that 3% of level 1 reviews and 3% of level 2 reviews were found to be not in order. This suggests that the consistency in the level of inaccuracy in this test site situation remains the same whether it is at level 1 or 2.

Reasons why MCCDs were rated as 'not in order'

2.14 Based on analysis^[6] of the 41 reviews where the MCCD was not in order and a replacement was requested, concern about the cause of death information was the most common reason for requesting a replacement MCCD. The MR identified an incorrect or incomplete cause of death in 36 (or 88%) of the 41 cases and an incorrect sequence of cause of death in 13 cases (32%). Other reasons include missing information about the certifying doctor (identified in 17 or 41% of the cases) and inaccurate or missing personal details about the deceased (identified in 11 or 27% of the cases). Further information about the concerns about these MCCDs is listed below.

• Incorrect or incomplete cause of death (36 cases)
  Cause of death being wrong or too vague
  Abbreviations being used for cause of death
  Conditions being omitted from the MCCD or being recorded when they should not be.
• Information about the certifying doctor being omitted (17 cases)
  Official address of certifying doctor missing
  Certifying doctor details omitted (e.g. name)
  Medical qualifications missing.
• Incorrect sequence of cause of death (13 cases)
  Sequence of cause of death being incorrect or illogical.
• Inaccurate or missing personal details about the deceased person (11 cases)
  Place of death missing
  Time/date of death missing or incorrect
  Deceased's personal details (e.g. name) wrong or omitted.

Quality scorecard ratings

2.15 MRs rated each MCCD they reviewed on a quality scorecard. Initially, MRs rated each MCCD on a scale of one to five on three elements: cause of death, referral to Procurator Fiscal and other information.

2.16 The results of this scoring show that there were few concerns about the accuracy or the quality of MCCDs reviewed. For example, MRs rated the vast majority of MCCDs (407, or 85% of all MCCDs scored in this way) as 'five' for cause of death, ie there were no concerns about the accuracy of the cause(s) of death listed on the MCCD.

2.17 Part way through the test sites, the quality scorecard system was changed. MRs were asked to rate the quality of seven aspects of each MCCD:

• the personal data on the deceased (on a scale of incorrect/missing, trivial error or correct);
• diseases or conditions relating directly to death (incorrect, questionable/imprecise or correct);
• other significant conditions contributing to death (incorrect, questionable/illogical, correct);
• tick boxes on MCCD (not used, incomplete, complete);
• Procurator Fiscal referral (not used, incomplete, complete);
• certifying doctor or responsible consultant information (incorrect/missing, trivial error, complete/complete); and
• overall legibility (very difficult to read, some difficulty to read, easy to read).

2.18 Similarly, these ratings showed that the majority of MCCDs had been completed accurately and fully. For example, 705 or 86% of the 822 MCCDs rated in this way were rated as 'correct' for diseases or conditions relating directly to death and 761 or 93% were rated as 'easy to read' in terms of overall legibility.

2.19 As noted above, the most common reason for rating an MCCD as not in order was a concern about the quality or accuracy of the information given about the cause of death. Concern about an incorrect or incomplete cause of death was noted in 36 of the 41 cases and an incorrect sequence of cause of
death was identified in 13 cases.

2.20 Taking into account all MCCDs reviewed, including those rated 'in order' and 'not in order', the areas of most concern arising from the quality scorecard ratings relate to:

• the provision of information about the certifying doctor and responsible consultant - this information was 'incorrect/missing' on 104 or 13% of the MCCDs reviewed using the new style scorecard and included a 'trivial error' in 50 cases (6%);
• information about causes of death - 20 cases (2%) were incorrect in terms of diseases or conditions relating directly to death and 97 (12%) were rated as questionable/imprecise. Similarly, 14 cases (2%) were rated as incorrect in terms of other significant conditions contributing to death and 62 (8%) were rated as questionable/illogical; and
• information about the deceased person - 53 cases (6%) were rated as incorrect/missing and 43 cases (5%) were rated as including a 'trivial error'.

2.21 This analysis again shows that the majority of MCCDs are completed fully and accurately but some doctors might benefit from some education and training related to completing MCCDs, particularly in terms of information about causes of death and details about the certifying doctor and the deceased person.

2.22 We provide full details of the quality scorecard ratings in Appendix 2.

Referrals to Procurator Fiscal

2.23 In the vast majority of reviews, MRs were satisfied that the case did not require referral to the Procurator Fiscal (PF). Combining the results of the new and old quality scorecards shows that MRs were fully satisfied that the case did not warrant a referral to the PF in 98% of reviews^[7]. In the 2% of cases where MRs considered a referral to the PF might be necessary, this was often done informally and there appears to have been no increased burden for the PF service.

Duration of reviews

2.24 Gathering accurate and complete data about the duration of reviews is an important element of this test site and evaluation as this will be a critical aspect of the process when it is implemented nationally. Initially, MRs provided very little data about the duration of reviews because they were uncertain about how to record this information. Consequently, NHS ISD discussed this issue with the test sites and provided further guidance on how to record duration.

2.25 Tables 2.4 and 2.5 display the data provided by test sites after NHS ISD had clarified how to record the duration of reviews. These tables show how many hours the MRs spent working on each review and how many days (including weekend and bank holidays) elapsed in total between receiving the referral and completing the review.

Table 2.4: Duration of reviews from referral to signing off of MCCD in Dundee

	Level 1 (Basic)				Level 2 (Comprehensive)				Upgraded from Level 1 to Level2				Total
	Hours working on MCCD form from referral to sign-off				Hours working on MCCD form from referral to sign-off				Hours working on MCCD form from referral to sign-off
Total number of days from referral to sign-off	Blank	Less than 1 hour	1 hour or more, but less than 3	3 or more hours	Blank	Less than 1 hour	1 hour or more, but less than 3	3 or more hours	Blank	Less than 1 hour	1 hour or more, but less than 3	3 or more hours
Blank	-	1	-	-	-	-	-	-	-	-	-	-	1
Less than 1 day	1	212	9	1	-	-	23	6	-	-	-	-	252
1 day or more, but less than 3	-	12	-	1	-	-	-	1	-	1	1	-	16
3 days or more, but less than 7	-	29	4	1	-	1	1	4	-	-	-	-	40
7 or more days	1	10	6	6	-	-	3	5	-	-	-	1	32
Total	2	264	19	9	0	1	27	16	-	1	1	1	341

2.26 Table 2.4 shows that NHS ISD received data about the duration of 341 reviews in Dundee.

• The majority (72%) of level 1 reviews were completed with less than one hour's work within 24 hours of the referral being made.
• Eight per cent of level 1 reviews took seven or more days to complete.
• Nearly half (49%) of level 2 reviews (including upgraded reviews) were completed with between one and three hours' work within 24 hours of the referral being made, but 13% took between three and seven days to complete and 19% took more than seven days to complete.

2.27 Thirty-two reviews in Dundee took longer than seven days to complete. NHS ISD provided further analysis of the circumstances surrounding these reviews. This analysis found that half (16 of the 32) were recorded as being concluded within seven or eight days. However, there were also four instances where the total time taken to complete the review was four weeks or more, with one case taking 130 days to sign off following its referral due to difficulties in contacting the certifying doctor, who had moved on to another placement, difficulties in accessing the medical records, which had been sent to a local authority storage site, and difficulties in contacting the responsible consultant.

2.28 This evidence shows that the MR in Dundee completed most referrals within the intended timescales for reviews, although there were some significant delays due mainly to obstacles related to contacting doctors and accessing medical records.

Table 2.5: Duration of reviews from referral to signing off of MCCD in Dumfries & Galloway

	Level 1 (Short)				Level 2 (Comprehensive)				Upgraded from Level 1 to Level2				Total
	Hours working on MCCD form from referral to sign-off				Hours working on MCCD form from referral to sign-off				Hours working on MCCD form from referral to sign-off
Total number of days from referral to sign-off	Blank	Less than 1 hour	1 hour or more, but less than 3	3 or more hours	Blank	Less than 1 hour	1 hour or more, but less than 3	3 or more hours	Blank	Less than 1 hour	1 hour or more, but less than 3	3 or more hours
Blank	-	114	88	102	-	8	12	22	-	-	-	2	348
Less than 1 day	3	1	-	1	-	-	-	-	1	-	-	-	6
1 day or more, but less than 3	2	1	-	-	1	-	-	-	-	-	-	-	4
3 days or more, but less than 7	22	-	-	1	14	-	-	1	3	-	-	-	41
7 or more days	2	-	-	2	4	-	-	-	1	-	-	-	9
Total	29	116	88	106	19	8	12	23	5	0	0	2	408

*Note: In 348 cases, the MR reported data related to the number of hours spent working on the review, but not related to the number of days that elapsed from referral to sign off.

2.29 Table 2.5 shows that NHS ISD received data about the duration of 408 reviews in Dumfries & Galloway. Just over a third (34%) of level 1 reviews were completed with less than one hour's work, 26% involved one to three hours' work and 31% involved more than three hours.

2.30 While this gives an indication of how much time the MR spent working on each review, it is not clear how many days passed between receiving the referral and completing the review because of a lack of clarity around the data recorded on the ISD reporting form.

2.31 However, the MR in Dumfries & Galloway carried out additional analysis of the duration of her reviews. Table 2.6 outlines the findings of this analysis. Please note that we have calculated the figures below using the median rather than mean average because there are a few cases where reviews took radically longer, and these outliers would artificially inflate the mean average. For example, one case took 624 hours (26 days) from referral to sign off.

Table 2.6: Duration of reviews in Dumfries & Galloway (hours)

Measure	Level 1 Reviews	Level 2 Reviews (including upgraded reviews)	All Reviews
Elapsed time from picking up referral to signing the MCCD off (median hours)	1.55	23.45	1.95
Hours working on MCCD from referral to sign off (median hours)[8]	0.50	1.43	0.50
Time spent reviewing medical records (median hours)	Not applicable	0.75	Not applicable
Time spent trying to contact doctors (median hours)	0.05	0.10	0.05
Time spent discussing cases with doctors (median hours)	0.05	0.10	0.05

2.32 This shows that in most cases, reviews are completed within the expected timescale but the duration of a review can be affected drastically by factors such as delays in contacting doctors or accessing medical records. One per cent of level 1 reviews in Dumfries & Galloway took seven days or longer to complete.

2.33 We discuss the factors that affect the duration of reviews more fully later in this chapter.

Summary of case study evidence

2.34 We analysed the process and outcome of 18 case study reviews. We asked the MRs to select case study reviews based on the criteria in Table 2.7.

Table 2.7: Case study criteria

Criteria	Target number	Number achieved - Dundee	Number achieved - Dumfries & Galloway	Number achieved - total
Level 1 reviews where the MCCD was not in order	2 (one in each area)	1	1	2
Level 2 reviews where the MCCD was not in order	2 (one in each area)	1	1	2
Upgraded reviews	2 (one in each area)	1	1	2
Level 1 reviews which went well	2 (one in each area)	1	1	2
Level 2 reviews which went well	2 (one in each area)	2	1	3
Upgraded reviews which went well	2 (one in each area)	0*	1	1
Level 1 reviews which were difficult	2 (one in each area)	2	1	3
Level 2 reviews which were difficult	2 (one in each area)	1	1	2
Upgraded reviews which were difficult	2 (one in each area)	0	1	1

Note: only one review had been upgraded in Dundee when the MR selected the case studies so she selected one additional level 1 and one level 2 review instead.

2.35 The summary below gives some insight into the findings of our case study analysis.

• Reviews seemed to get more efficient as the test sites progressed - MRs reported being able to complete reviews more quickly as they became more used to the process, although delays in accessing doctors and medical records continued to lengthen the duration of some reviews.
• In three cases we analysed, MRs were not able to speak to the certifying doctor due to him/her being on leave or having moved out of the area.
• The activities associated with a review take a short time to complete but delays in contacting certifying doctors and/or medical records can increase the duration of a review significantly. In all case studies, the MRs' conversation with the certifying doctor (or another medical professional if the certifying doctor was not available) took only a short time, ranging mostly from three to 15 minutes with one example of a longer discussion of 37 minutes. For example, one level 1 case study review took 91 days from referral to sign off because the MR could not contact the certifying doctor and had difficulty in identifying another member of the medical team who could discuss the case. Eventually she spoke to a senior consultant and completed the review with a 15 minute conversation. In another case, the MR tried to track down medical records that had gone missing. The delay caused by this meant the review took 23 days to complete.
• We reviewed six cases where a replacement MCCD was issued, mostly because of inaccuracies around the causes of death. In addition to these cases, the MRs raised concern about the following aspects of the MCCD:
  • legibility (two cases);
  • missing information about the certifying doctor (two);
  • diseases and conditions directly leading to death (two);
  • other significant conditions contributing to death (one);
  • incorrect or missing information about the deceased (one); and
  • debatable Procurator Fiscal referral (one).

The processes within the test sites

2.36 Below we present qualitative findings relating to the process within the test sites, organised under the following headings:

• receiving referrals
• upgrading reviews
• factors that affect the duration of reviews
• faxing referrals back
• Medical Reviewers' workload
• outcome of reviews
• reporting to ISD
• location of Medical Reviewers
• professional background of Medical Reviewers
• liaising with doctors
• registrars
• bereavement support officers.

Receiving referrals

2.37 In the test sites, all referrals were sent to the MRs by local registrars using a random sampling method.

2.38 Most registrars faxed referral documents to the MR. This created problems including:

• missing information - in some instances, the fax cut off parts of the documents;
• legibility - the quality of the fax transmission can affect the legibility of the documents; and
• confirming receipt - there is no confirmation that a fax has been received by the intended recipient so in Dundee registrars emailed the MRA to let her know when a fax had been sent, and the MRA replied to confirm receipt. This provided extra work for both parties.

2.39 MRs and MRAs in both areas agreed it would be easier if documents could be emailed to MRs. Two registrar offices in Dumfries & Galloway took this approach and this seemed to work well. However, some registrar offices may not have access to the equipment required to scan and email documents.

2.40 MRs and MRAs commented that there were often daily fluctuations in the number of referrals received which could lead to backlogs. MRs commented that they would be more able to avoid backlogs and complete reviews quickly if they received a steady flow of referrals from registrars.

2.41 When the process is implemented nationally, MRs will receive referrals from registrars using a random sampling method, but members of the public and other interested parties will also be able to request reviews from the MR. This element of the process has not been tested by these test sites.

Upgrading reviews

2.42 MRs had the ability to take cases referred to them as level 1 reviews and upgrade them to level 2 reviews. This did not happen very often - only 1% of cases were upgraded. Where cases were upgraded, MRs did so because they required access to the deceased person's medical records. The main reason for this was where the MR wanted to clarify details or potential inaccuracies on the MCCD but was unable to because the certifying doctor was unavailable and no other member of the medical team was able or available to answer the MR's queries. In some other instances, cases were upgraded when the MR's review of the MCCD and/or discussion with the doctor led her to conclude that the case might require discussion with the Procurator Fiscal (PF), and she required access to medical records to inform her decision about contacting the PF and to inform any discussions with the PF.

Factors that affect the duration of reviews

2.43 MRs commented that reviews can normally be completed within the intended time frame but the experience of these test sites shows that there are various factors which can hinder this and lengthen the duration of a review.

2.44 Firstly, the length of time it takes to contact and hold a conversation with the certifying doctor and/or the responsible consultant affects the duration of a review. A review can be completed promptly if a doctor is available quickly but this is not always the case because doctors: frequently work unusual shift patterns; may be too busy to be interrupted; or, especially in the case of junior doctors, may have moved on to work elsewhere, perhaps in a different health board area.

2.45 MRs in both areas noted the importance of being flexible in the times of day during which they contact doctors. Our case studies provide examples of the various times of day that MRs contacted doctors, and the MRs gave further anecdotes of speaking to certifying doctors outside of office hours in order to fit in with the doctor's shift patterns. Both MRs also commented on the importance of having access to doctors' rotas so they can identify times that the doctor is likely to be available.

2.46 MRs also commented on the importance of being flexible in the way in which they approached doctors. In some cases, it was easier to go to see doctors face-to-face, particularly if it was proving difficult to contact him/her by phone.

2.47 Secondly, the length of time it takes to access medical records affects the duration of a review. This can be problematic for a number of reasons:

• the records may be held at a hospital or surgery that is far away from the MR's base, meaning they may have to travel to see them or wait for a delivery from the NHS internal mail service;
• a person's records may be spread across more than one location, and this can add to the above difficulty;
• GP practices' medical records are held electronically but cannot be sent electronically outside the practice; and
• MRs and MRAs in the test sites have found that Topas, the database which records where records are stored, is sometimes inaccurate.

2.48 This issue affected both test sites but was particularly prominent in the early stages of the test site in Dumfries & Galloway as a result of the rural nature of the area. Initially, the MR went to look at medical records in person and, given the size of Dumfries & Galloway, there were instances when the MR spent significant time driving long distances to review records held at GP practices or hospitals outwith Dumfries. This also involved a preparatory telephone call to the practice or hospital to arrange a suitable time for the MR to visit. The MR overcame this by asking GPs to fax a four page summary (known as the intermediate/emergency care summary) of the records to her and by making use of the NHS internal mail system to access records from community hospitals^[10].

2.49 To facilitate the prompt completion of reviews, MRs and MRAs commented that it would be helpful if they could access medical records electronically from their base.

Faxing referrals back

2.50 The MR in Dumfries & Galloway noted that there is sometimes a delay between signing the MCCD off and faxing the documents back to the registrar. The median figure for this delay is only three minutes, but 11% of reviews involved a more significant delay of 3.5 hours or longer. This is due to several reasons including instances where the MR completed the review at home but had to wait to get to the office to fax it back, or instances where the MR completed the review while the MRA was not in the office, and waited until the MRA returned to fax it back. The MR commented that it helps if the MR and MRA are located in the same office - initially the MR was based two floors away from the MRA but soon moved to the MRA's base where the fax machine was situated.

Medical Reviewers' workload

2.51 Medical Reviewers and Medical Reviewer Assistants in both areas commented on the heavy workload involved in the test sites. Both MRs reported working beyond their contracted 0.5 FTE hours to complete the required volume of reviews.

2.52 The workload of MRs and MRAs will be an important consideration when the process is implemented nationally. However, we should note that MRs in the test sites were asked to complete reviews at a faster rate than will be expected of MRs when the process is implemented nationally. The target of 1190 reviews for the test sites was set to ensure the MRs carry out a robust sample of reviews to allow lessons to be learned and conclusions to be drawn before national implementation. This means that MRs in the test sites were expected to complete around 15 to 16 level 1 and one to two level 2 reviews per week; but when the system is implemented nationally, MRs will be expected to complete ten level 1 and one level 2 reviews per week.

Outcome of reviews

2.53 The vast majority of reviews found MCCDs to be in order but there are some examples where reviews highlighted areas that doctors may require some guidance on. For example, where MCCDs were found to be not in order, this was most commonly due to concerns about information given about the cause of death. Considering all MCCDs, including those rated as 'in order' and 'not in order', where reviews found any concerns about the MCCD's quality or accuracy, these concerns related mostly to issues such as cause of death information, legibility and doctors failing to include required information such as the place of death or their GMC number.

2.54 There appears to have been slight differences in approach between the MRs in Dundee and Dumfries & Galloway. The MR in Dundee asked for more replacement MCCDs than in Dumfries & Galloway. This illustrates the imoprtance of consistency and will be commented on further in the final chapter.

2.55 Both areas introduced a system whereby, if the MR had minor concerns about an MCCD but did not want to issue a replacement MCCD, she would type a letter outlining the changes, ask the certifying doctor to sign it and then ask the registrar to append the letter to the existing MCCD.

Reporting to ISD

2.56 When MRs completed a review, they completed a form with details of the review which they then submitted to NHS ISD. This provided information about the process and outcome of each review.

2.57 The form received various revisions over the course of the test site to ensure it was easy to complete and captured the necessary information. For example, the quality scorecard was revised with the benefit of providing an indication of which element(s), if any, of the MCCD were of concern.

2.58 In both test sites, the MR completed the MR form by hand and the MRA typed it up. This seems a labour intensive process and could be simplified.

2.59 There was some uncertainty among MRs and MRAs about the information they were expected to include on the form. For example, there was some uncertainty about how to record the duration of review. Initially, MRs were unsure whether to record the total time that elapsed from referral to completion of review, or just the time spent working on the review. NHS ISD discussed this issue with the MRs and provided advice on the required information. This illustrates the importance of providing clear guidance to MRs and MRAs about the information they are expected to provide on the reporting form.

Location of Medical Reviewers

2.60 The MR in Dumfries & Galloway benefitted from being based at the Dumfries & Galloway Royal Infirmary: this meant she could easily access stored medical records as well as doctors based at the hospital, and receive deliveries of medical records from community hospitals via the NHS internal mail service. On the other hand, the MR in Dundee was based at Dundee University. This caused complications in terms of:

• setting up IT systems - there was a delay at the beginning of the test site while arrangements were made for the MR to access NHS IT systems from her base;
• accessing medical records, most notably because the NHS internal mail system in Dundee would not deliver medical records to non-NHS sites so the MR had to go in person to view records; and
• contacting doctors - if the MR wanted to speak to a doctor face-to-face, she had to drive to the doctor's place of work to do so.

2.61 The issue of the MRs' location and the ease with which they can access NHS IT systems, medical staff and records will be an important consideration when the process is implemented nationally.

Professional background of Medical Reviewers

2.62 The professional background of MRs is another important consideration. Both MRs in the test sites benefitted from being medical professionals who have worked in their area and who are well-known among doctors in the area. This helped them to liaise with doctors because they have credibility and, in some cases, established relationships with doctors in the area. This helped to make doctors more receptive to taking part in the review process.

2.63 In addition, MRs found their knowledge of local personnel helpful. For example, the MR in Dumfries & Galloway commented that knowing the name of a senior doctor's secretary can make it easier to arrange an appointment with the doctor.

Liaising with doctors

2.64 MRs reported that, in general, doctors were happy to take part in the review process. Moreover, the doctors who we spoke to as part of case studies were generally happy to take part in the evaluation and junior doctors commented on the learning opportunity that discussing the MCCD with the MR provided.

2.65 The test sites found that it was easier for the MRs, rather than MRAs, to make initial contact with doctors. The MRs were more likely to receive a positive response from doctors given the credibility they had as a result of their professional background. This approach also reduced the burden on doctors: rather than being disturbed twice (firstly with a call to arrange the appointment and then a phone call or meeting to talk to the MR), their involvement in the review could be finalised within one phone call.

Registrars

2.66 The new process creates a small amount of extra work for registrars to prepare documents for referral and to file away the documents after completion of the review. In addition, where a replacement MCCD is issued, the registrar must send this to National Records Scotland and change the cause of death listed in the register.

2.67 Another issue related to the involvement of registrars in the process is their responsibility to tell bereaved people about the new process and that it has the potential to delay funerals. Registrars reported little reaction from bereaved people during the test sites when they told them about the review process, but registrars speculated that this could change when the process is live and has the potential to delay funerals.

Bereavement Support Officers

2.68 Bereavement Support Officers at Ninewells Hospital played an important role in the Dundee test site, particularly in following up cases where doctors had failed to complete the revised MCCD form required for the test site and in ensuring that hospital wards had a sufficient stock of the forms. This will not be an issue when the process is implemented nationally because there will only be one MCCD form to complete, but it is possible that Bereavement Support Officers and other similar staff could play a role in supporting the review process.

2.69 The next chapter of the report explores the findings from the evidence and stakeholder interviews around a number of themes.