Within hospital studies and adverse incident reporting, medication omissions are portrayed as a significant systemic patient care and service delivery issue. This study suggests that medication omissions are common within RAC homes, with 73% of residents having at least one dose omission during the period. However, the rate of omissions per 100 doses is low, and varies between levels of care.
This is the first study that reports on a large sample and provides robust and specific detail on medication omission rates. Hospital-based studies have found an omission rate per medication administration episode of up to 11%, also stating that one in nine doses were omitted. This study has found that only 2.93% of dispensed medication doses were omitted over the timeframe, significantly fewer than indicated by previous studies. This is a significant finding, as RAC homes are the major site for medication administration to patients/residents outside of hospital ward settings. Hospital-based studies typically contain a large proportion of new admissions and refer to medication not being available as a key contributor to medication omissions. We would expect the medication omission rate to be less in RAC homes as they do not have the high volume of admissions, and individuals typically stay in RAC longer than in a hospital setting. This study sets out a clear mean omission rate per resident, 3.59 doses/100 doses (s.d. 7.43), a level of detail which previous studies have been unable to provide.
Medication Omissions & RAC Ownership
The rapid ageing of the New Zealand population has led to an increase in large organizations and corporate entities governing and building RAC homes. Corporate-owned RAC homes make up 66% of the sample for this study, and prior studies have not considered ownership types in relation to medication omissions. Corporate-owned RAC homes had a slightly higher rate of omission, both in general, and for medication omissions classified as not-administered (Table 3.) This may reflect a lack of a clear definition for the not-administered category within these RAC homes.
Reviews suggest that corporate entities and for-profit RAC homes may provide lower quality care than independent or non-profit sites. While research mainly focussed on the United States sector , studies in Europe also show increased signs of neglect in for-profit facilities . In New Zealand each independent RAC home is responsible for its own policies around medication administration and management, whereas corporate entities operate under a standard policy written by the wider organization. It is unclear whether differences in omission rates between ownership types reflect variances in quality of care, or policies, and as such, we cannot justify omissions as a quality indicator, however our findings are consistent with international research suggesting variation in outcome related to ownership type of RAC homes.
Categories of Omission & Recording Practices
Refusals and medication unavailability have commonly been identified as the top two categories/reasons for omission[9, 10]. Refusals indicate a resident’s decision not to take an offered dose. The reasons behind this can relate to residents exercising autonomy, their preferences; on the other hand, there may be no clear rationale, and the resident may refuse medication for reasons that care staff cannot ascertain. The best practice guidelines prompt care staff to attempt administration up to three times, and to record a reason for the refusal. The relationship between omissions and errors requires more consideration, as a resident refusal does not imply a clinical error, despite resulting in a dose omission. Almost 35% of omissions were refusals, significantly higher than prior studies, with a mean rate of 1.27 doses per 100 dispensed doses/resident . Higher refusal rates were found in palliative, transitional, and dementia care types. Refusals of medication may be accepted more in these settings—care staff may persist with administration more in residential and hospital settings.
Medications ‘withheld’ based on care staff decision-making were the least frequent category of omission recorded, with 15.5% of all omissions. The withheld medication rate per 100 dispensed doses was consistently between 0.29 and 0.59 (s.d. 1.28-1.64) doses, with higher rates in Dementia Hospital and Palliative Care (Table 2.) However, widespread inconsistent recording of why medications were omitted hinders our ability to explain why these medications were withheld. Care staff have an obligation to record clinical decisions and reasons for omission, so that prescribers, other staff, and the residents themselves can access these reasons. Yet New Zealand’s best practice guidelines provide less guidance on how to make decisions and withhold medications, compared to resident refusals[3, 4]. This implies that omissions based on clinical decisions are not subject to the same level of scrutiny that resident-initiated omissions are. There is, however a fine balance between processes and improving safety culture . Our study adds to the international literature on medication errors and omissions [16, 17].
The most prevalent category of omission was ‘not administered’, a category originally designed to indicate medication or resident unavailability. Prior studies have used up to twelve different categories of omission, many of which can be grouped under the three categories present in Medi-MapTM. For example, ‘vomiting’ has been used as its own omission category, but in Medi-MapTM this falls under ‘withheld’, as this is typically a reason for staff to initiate an omission. However, the use of fewer categories requires staff to better record why an omission has occurred, as ‘not-administered’ is not enough clinically to justify the omission; the care team also need to know why it was not-administered. The use of a less bounded category by the data provider for this study allows us to highlight the need for clear definitions and the provision of a list of common reasons for not-administering a medication. Of all omissions, 50% were recorded as ‘not administered’, with the highest omission rates per 100 doses, particularly related to Rest Home Care, Dementia Residential, Palliative, YPD and Transitional Care.
There is no real way of knowing why these medications were not given. Supply issues have been highlighted as complex to fix. The broadness and frequent use of the ‘not administered’ category indicates that staff may not be following up in situations where medication was not given. There are a lack of comments explaining why a medication has been ‘not-administered’, and if a medication was omitted then staff returned to administer it later, then this medication dose would have shifted to ‘administered’ and not been present in our dataset. The ‘not-administered’ category has been adopted as a ‘catch-all’—comments that are left refer to medications being withheld or refused, despite these being their own separate categories of omission. Providing a menu of common reasons alongside an option to record a novel reason may be useful within e-records and would save time from a care provision perspective. Failure to record this information can obstruct the provision of care, leading to miscommunication, heightened risk to residents, delays in ordering or following up prescribed medications. This is a case where technology can support enhanced professional practice and increase the quality of clinical care, if it is used to its full potential.
This study confirms that laxatives and mild pain relief are the most commonly omitted medications during medication administration. This is in line with prior studies of medication omissions and errors in palliative, RAC home, and hospital settings[5, 9]. The omission of specific medications in RAC homes in New Zealand will be further discussed in a subsequent paper, as they are an additional concern from a quality perspective.
Use of Electronic Medication Records to Track Omissions
Electronic medication records, like their paper predecessors, are only as good as the user. They facilitate large scale analysis of long-term medication and administration data but require consistent recording and adequate detailing. There appears to be a relationship between ownership of RAC homes and how omissions are recorded from a rate of omission and omissions type perspective. There also appears to be a relationship between a resident’s care level and rate of omissions. This indicates that organisational policy and the type of care a resident is receiving may have an influence on care staff’s decision-making when a medication has the potential to be omitted.
A review of the medication administration guidelines for RAC homes is needed, both within New Zealand and internationally, considering the widespread adoption and potential of e-records. In New Zealand, records tell us that between the years of 2016-2017, 72.7% of residents within the RAC home sample for this study experienced a medication omission. Yet the best practice guidelines barely acknowledge omissions as a concern during medication administration—except for refusals. Omissions are listed as a form of error, situated beside medications given at the wrong dose time [3, 4, 11]. Internationally, research into care staff perceptions of medication omissions is needed, as well as a review of how omissions are treated in Best Practice Guidelines. Further work on supply chain , repeat omissions , and organisational culture related to medication use would expand this area.
There were three predominant limitations to this study. First, that the secondary data only tracked the care level funding of residents upon admission to their care home, rather than care level changes over time. The needs of RAC Home residents are likely to increase over time, and care levels and funding naturally change to meet these needs over time. Second, the data collection parameters did not include data on why medications had been prescribed, limiting our ability to discuss why certain medications were prescribed more frequently than others. Third, the data collected only related to dispensing and administration of medication—as dispensing data, RAC home, and resident data were used as denominators we cannot be certain that the proportion of medication omissions is not an underestimation. An analysis of prescribing data would help to clarify this, as this study focused upon events at a RAC home level.