Consensus on quality indicators to assess the organisation of palliative cancer and dementia care applicable across national healthcare systems and selected by international experts

The IMPACT consortium consists of experts of the pan-European research group on detection and timely INTERvention in DEMentia (Interdem) and the European Association for Palliative Care (EAPC), all of whom are stakeholders in their country in palliative cancer and/or dementia care. The research team was invited to use their networks to purposefully select panellists for the modified RAND Delphi procedure. A key selection criteria was that the expert had to have extensive knowledge about palliative care, cancer care or dementia care. Additionally, experts had to be able to communicate in English (both verbally and non-verbally) and were planning to attend the 2012 Congress of the European Association of Palliative Care (EAPC) in Trondheim, Norway. Project partners nominated national and international experts in palliative cancer and dementia care. All nominated experts were approached via email (n = 50), forty experts from twelve countries agreed to participate in the modified RAND Delphi procedure (Table 1). All participants provided written informed consent. About half of them were experts in palliative cancer care and the other half in dementia care. Twenty-two panellists were active clinicians in this field (e.g. physician, nurse, psychologist, etc. currently involved in direct-patient care), the others were researchers.

A search for existing sets of QIs was conducted in PubMed. The search strategy was limited to English literature and consisted of various search terms that referred to subject-specific keywords describing palliative care (combined using “or”), as well as (“and”) the assessment of care using QIs (combined using “or”). Synonyms and medical subheading terms were used to fully include relevant literature (see Table 2).

Subsequently, references of key papers describing sets of QIs were hand searched. Additionally, consortium members of the IMPACT project were asked to nominate national and international sets of QIs on palliative cancer and dementia care they considered important. Two researchers (YE & JvRP) subsequently reviewed all of the identified QIs independently to determine if the QIs assessed the structure and process of palliative care and to structure them according to the domains of the recommended framework for the organisation of palliative care of the Council of Europe [21].

For the first written Delphi-round (April 2012), panellists received a personal invitation for an online questionnaire. To reduce the large number of identified QIs, panellists were asked to nominate one QI per domain of palliative care [22]. Those QIs that were nominated by the panellists were included in the second round of the modified RAND Delphi procedure.

In the second round (May 2012), also via an online questionnaire, panellists were asked to rate QIs on a 9-point Likert scale for clarity (1 = not clear at all; 9 = very clear), usefulness (1 = not useful at all; 9 = very useful), to rephrase unclear and to add missing QIs. They were instructed to rate a QI high on usefulness if it: 1) corresponded with a basic quality level; 2) referred to a higher quality level that would be met only in very good practices; or 3) was associated with an innovative quality level which is exceptional at the moment, but could become the optimal quality level in the near future [23]. They were asked to give a low rating on usefulness if a proposed QI: 1) was too ambiguous or represented an unrealistically high quality level; 2) did not correspond with the material, social or cultural conditions of the situation in their country; or 3) was not in accordance with the regulations of palliative cancer and dementia care in their country [23].

A consensus meeting was organised during the EAPC Congress June 7, 2012 in Trondheim, Norway. The meeting was chaired by an independent researcher with the aim to reach consensus on the QIs on which there was disagreement or where the median score was between 4 and 6 in the second Delphi round. Participants received a rating sheet on which the median ratings of the second Delphi-round of all experts were visible. Participants were given 30 minutes to rate the adapted QIs for usefulness. Next, per QI, participants were asked to raise their hand if they had rated usefulness 6 or less. If at least nine (30%) participants raised their hand, [20] the QI was discussed until consensus was reached.

After the consensus-round, the remaining indicators were fed back to the panellist with the purpose to validate the changes that were made (September 2012).

In the final step of the QI development process, QIs were operationalised by the research team into questions that could be used by healthcare professionals to assess their organisation of palliative care and identify areas for improvement. During this process, it appeared that some QIs were inappropriate or not measurable (e.g. too time consuming to answer them appropriately). All QIs were therefore rated for necessity by the IMPACT research team (October 2012), representing both clinicians and researchers that also took part in the modified RAND Delphi procedure. QIs that were considered not necessary after this round, were omitted from the list.

QIs with a median rating on the usefulness scale of 7, 8 or 9 without disagreement were considered to have face validity. Disagreement was defined as: 30% or more of the panellists rated a single QI in the 1–3 tertile and more than 30% in the 7–9 tertile. QIs scored with a median of 1–3 without disagreement were not considered to have face validity. Because panellists had rated QIs high on usefulness, only QIs with median ratings of 8 or 9 were considered face valid for the second Delphi-round. Only QIs that were rated valid by all panellists were included in the final set [20].

The Medical Ethics Committee of the district Arnhem-Nijmegen has declared that this study doesn’t fall within the remit of the Medical Research Involving Human Subjects Act (WMO) (registration number 2012/075). This means that this study can be carried out without an approval by an accredited medical ethics committee.

The availability of a dedicated palliative care team was considered important by almost all panellists in the Delphi procedure. They explicitly stated that palliative care services should not only be available during office hours, but at all times (day, evening, night and weekends). Furthermore, specific elements of palliative care were considered important, such as the availability of opioids and anticipatory medications for symptom control, as well as the availability of bereavement support.

Almost all panellists also rated important the accessibility of the medical record to health care professionals, timely transfer of information between settings (including when transferring or discharging patients). Panellists also considered an assigned contact person, who maintains regular contact with patients and their families, useful.

The infrastructure of the place where palliative care is provided, such as a single bed hospital room, was rated important. Access to equipment (such as anti-decubitus mattresses, suction equipment, etc.), required to provide palliative care, was considered important. Panellists also rated high consensus for facilities for relatives to visit, stay overnight, and a private area for saying goodbye to the deceased. There was no agreement on QIs that aimed to control waiting time or waiting list, i.e. these aspects were not considered to be unique for palliative care and therefore not important.

Regular assessment of pain and other symptoms was rated as a valid quality criterion, though it was commented that a validated instrument might not always be available, particularly for specific patient groups (e.g. for persons with advanced dementia).

There was agreement on the need for a multidisciplinary team, which should consist of at least a physician and nurse, and have access to a range of supporting disciplines, such as: physiotherapist, psychologist, occupational therapist, social worker, dietician, and chaplain. Panellists also rated a regular multidisciplinary team meeting important.

Panellists recognized the importance of having a well-structured medical record. However, a QI about the structure of the medical record was not considered important by the panellists. Panellists only considered the inclusion of a medication regimen in the medical record important. They also considered a timely assessment (within 48 hours) of pain and other symptoms, psychosocial and spiritual needs, patient preferences, wishes and needs, and the patient's capacity to be involved in the decision making process as important.

Furthermore, almost all panellists rated the documentation of communication on the medical condition, goals of treatment, physical, psychosocial, and spiritual needs of the patient and their relatives, intention to return home, advanced directive, and end-of-life decisions as important.

Panellists rated the QI about assessing the experiences of care givers with the palliative care service important. Secondly, a QI about the quality of care, assessing the use of an end-of-life care pathway within the last three days of life, was also considered useful.

QIs about the staff’s learning objectives and a program for specialised and/or continuing medical education about the physical, psychosocial, and spiritual needs of a patient in need for palliative care were not rated important. There was also no agreement on disease-specific education for staff members, but panellists considered palliative care training specified to the professiona’ls background important.

Strengths of this study were that we used a large international group of panellists for our Delphi procedure, who were actively involved in palliative care (such as members of the European Association for Palliative Care). Secondly, by organising our consensus meeting during the EAPC Congress in Trondheim, Norway, key persons active in the fields of palliative care and dementia care were able to contribute extensively to the discussion of the Delphi procedure. Thirdly, the multidisciplinary character of palliative care was represented by the panellists (e.g. physicians, nurses, psychologists, etc.) involved in the Delphi procedure. Furthermore, half of them were professionals active in dementia care. Fourth, combining QIs for the organisation of services that provide care to palliative patients with cancer and those that provide such care to patients with dementia is unique. Our QIs can therefore be used in different settings.

A limitation of this study is that this set of QIs might not be comprehensive. Because an international, generic set of QIs was developed, some QIs that were important in only one or a few countries were excluded from the list. For that reason, important national or setting-specific QIs must be added when the set is used in a specific country. Secondly, this set of QIs is only related to the organisation of palliative care. Outcome and patient-related outcome measures were not included because they address a distinct purpose in measuring quality of palliative care. Thirdly, participants of the modified RAND Delphi procedure were selected because of their knowledge about palliative care, cancer care or dementia care. Because some experts of two large European networks (EAPC and InterDem) are part of the IMPACT consortium, they were also selected as participant for the modified RAND Delphi procedure (n = 18). Although not all countries were represented (like France) and others were overrepresented (like the Netherlands), the experts covered 12 countries from three continents, covering different health care systems and types of organisation of palliative care. Pilot testing the set of QIs in those countries and continents that were not represented in this study will reveal whether they are applicable in these countries too. Fourthly, unfortunately, there were no patient representatives involved as panellist. Testing the final set of QIs will therefore also have to incorporate their views on the basic quality level or higher quality level that would be met only in very good services.