Article Text
Abstract
Objective The Lund Integrated Medicines Management (LIMM) model improves the patient medication process and reduces primary care contact and rehospitalisation. The objective was to evaluate the quality of medication management activities and the time spent on these activities using the LIMM model in hospital and primary care.
Methods Questionnaires were distributed to physicians and nurses in hospitals, with and without the LIMM model, and in primary care. A time study of the activities of clinical pharmacists was also performed.
Results Responses were received from 67 physicians and nurses working in hospitals and 210 in primary care. The respondents thought that the quality of medication management would be much improved using the LIMM model. The model was associated with total median time savings by nurses and physicians of at least 1 h per patient, while the clinical pharmacist spent only 1 h with each patient.
Conclusion The LIMM model reduced the total time required for each patient by at least 1 h and improved the quality of the process.
Statistics from Altmetric.com
Introduction
Drug errors and other drug-related problems (DRPs) are commonly associated with patient injury.1 ,2 The Swedish government estimates that 3000 people die every year as a result of DRPs and that 6–16% of hospital admissions are drug related.3 The cost of avoidable medication-related injuries is estimated at €0.6–2.7 billion/year.
There is evidence that clinical pharmacy interventions reduce unnecessary costs.4 The Lund Integrated Medicines Management (LIMM) model5 includes systematic activities for medication reconciliation and review, education for the care team, and information and guidelines for communication with the patient to identify, resolve and prevent DRPs. In brief, at admission the pharmacist prepares a current medication list and identifies patient problems in the areas of knowledge, practical management, attitudes and adherence. The nurse evaluates the patient's symptoms and the pharmacist carries out a medication review. Subsequently, the team prepares a care plan based on identified DRPs. At discharge, the physician writes up the discharge information (LIMM-DI) containing a current medication list and a medication report on any changes. This LIMM-DI is given to the patient and sent to their primary care representative. The LIMM model, which has been investigated in four PhD theses and 14 studies, improves the appropriateness of medications, reduces the number of transmission errors and halves the number of medication-related contacts with healthcare professionals.5,–,10 The pharmacist's role and actions are seen as important.6 ,10
Aims of the study
The aims of the study were to investigate the estimated time spent by physicians and nurses on the medication process and the quality of the process with and without the support of the LIMM model. The time spent by pharmacists on LIMM-based activities was also investigated.
Methods
The questionnaires were tested on small groups before being used in the study. The quality of the LIMM-based medication process versus standard care was evaluated on a five-point scale: much better, somewhat better, no improvement, somewhat worse, much worse. Time utilisation was evaluated by calculating the difference in time spent on work activities with and without LIMM support and by direct questioning of healthcare professionals' perceptions.
The hospital questionnaires focused on medication-based activities performed by physicians and nurses at the Lund unit of Skåne University Hospital, where the LIMM model has been used since 2006, and a clinic in the Malmö unit of the same hospital (control).
The primary care nurse questionnaire focused on the distribution of drugs to patients after their discharge from hospital. Standard care, without information from the hospital, was compared with support from LIMM-DIs that were correct or contained one deliberate discrepancy. The nurses were asked who they contacted to clarify any problems. The general practitioner (GP) questionnaire asked how often they were contacted by the nurse to clarify problems.
The number of interventions and the time spent with each patient were recorded for four pharmacists performing LIMM-based activities on seven wards in Lund Hospital during a week in March 2010. The findings were compared with the number of recorded pharmacy interventions, and work rosters, for the fourth quarter of 2010.
Data are presented as medians and quartiles, and the paired t test was used.
Results
Responses to questionnaires were received from 12 (100%) physicians and 16 (100%) nurses at Lund and 12 (20%) physicians and 27 (28%) nurses at Malmö. In primary care, responses were received from 77 (78%) GPs and 133 (62%) community nurses.
Table 1 summarises the estimated time hospital physicians and nurses spend on each activity. At Malmö, physicians and nurses estimated that they spent 48 min per patient at admission and during hospitalisation (preparing a current drugs list and medication review). Most thought that LIMM support would be an improvement (much better 60%; somewhat better 37%). The physicians thought that these activities should take a maximum of 23 min, while the nurses thought 26 min. At Lund, all the doctors and nurses thought that the pharmacists' activities were important and should continue, and estimated that they saved 69 min at admission and during hospitalisation. The doctors thought the process should take 75 min at most, and the nurses said 50 min.
At Lund, all respondents stated that the LIMM-DI was time saving and they estimated that it would save them 53 min at discharge and after discharge (vs 89 min at Malmö).
It was calculated that the GPs would save 10 min using the LIMM-DI (n=61, range 5.5–15 min, p=0014 vs standard care) if they did not contact the hospital and 20 min if they did contact the hospital (n=11; range 15–26 min; p=0.002). On direct questioning, GPs thought they would save 15 min (n=68; range 10–30 min) regardless of whether the patient had no community care help, had community care help with medication or lived in a nursing home. Only 11 of 72 GPs (15%) chose to contact the hospital to seek help with discrepancies. These were the same GPs who indicated that they saved the most time with the LIMM-DI. When the LIMM-DI did not have discrepancies, the nurse did not need to contact the GP, and no extra GP time was needed. With one discrepancy, the GPs spent 14 min on each patient. Without the LIMM-DI, they spent 22 min on each patient.
It was calculated that the LIMM-DI would save the community nurses 21 min (n=126, range 3.0–31 min, p<0.001). On direct questioning, they thought they had saved 30 min (n=116, range 20–52 min) if the patient's LIMM-DI had one discrepancy and 39 min (n=124, range 30–60 min) if the LIMM-DI was correct.
The proportion of GPs who relied on the medication list increased from 11% to 48% when a LIMM-DI was issued; 72% stated that the LIMM-DI was much better and 23% that it was somewhat better. The corresponding results for community nurses were 71% and 27%.
During the 3 months of the study, 1126 patients were admitted to seven wards with LIMM pharmacists. The time spent per patient based on the quarterly statistics (60 min) and the time survey (65 min) corresponded very well.
Discussion
This study indicates that LIMM-based pharmacist activities take 1 h per patient. Physicians and nurses at hospitals and in primary care indicated that the LIMM model improves the quality of the patients' medication process. In addition, the model resulted in major time savings for physicians and nurses (of at least 2 h per patient) in all the areas listed in table 2. Additional primary care time savings can be expected if the LIMM-DI is accurate, that is, it has been reviewed by a pharmacist.6 ,7 The LIMM model also reduces the number of drug-related healthcare contacts and rehospitalisations, and appears to be cost effective.5 ,8 ,9
Of the GPs, 85% chose not to contact the hospital to seek help with errors in the medication lists provided at discharge when these were queried by the nurse. The consequences of this were not investigated in this study, but previous work has shown that there is an average of two errors per medication list at every discharge, and that this is reliably reduced to one when the LIMM medication report is issued.5,–,8 The problem partly stems from errors in the drug lists at admission, and partly from changes during the hospital stay not being documented or communicated to the patient and subsequent care providers. The LIMM model solves this by the pharmacist preparing an accurate medication list for the physician at admission and the discharging physician preparing a LIMM medication report of the changes made. The physician and nurse in primary care can then compare the patient's drug list prior to admission with that at discharge, with discrepancies between the lists reported in the LIMM medication report. We introduced a deliberate discrepancy in LIMM-DIs to highlight the improvement in quality when the LIMM model is used. The LIMM-DI decreased the total time spent by GPs and nurses by 29 min if there was one discrepancy and 52 min if the LIMM-DI was correct (table 2).
The unexpectedly large time savings in hospitals may have been the result of being studied, that is, the Hawthorn effect, although we tried to minimise this by using two methods: estimated and calculated time saved (table 2). The calculated results in primary care were obtained from the differences between LIMM reports that were correct and those with one discrepancy, while those in hospital care were derived from the two very similar units of the same clinic, from two cities. The low response rate from Malmö, in contrast to the 100% response rate at Lund, limits the study. However, this low response rate is not considered critical because the estimated times and time savings were similar between the units, indicating that the responses from Malmö were representative of the entire clinic.
Key messages
There is a need for evidence-based clinical pharmacy service models.
The LIMM model has been applied in hospitals and has proven benefits for the hospital medication process and for outpatient outcomes.
This study shows that the LIMM model improves the quality of medication processes and decreases the associated total time by at least 1 h for each admitted patient.
Conclusion
The LIMM model improves the quality of medication processes and saves at least 1 h per patient in hospital and primary care. These advantages and previously identified benefits will be the basis of a future pharmacoeconomics study.
Footnotes
-
Competing interests None.
-
Patient consent Obtained.
-
Provenance and peer review Not commissioned; externally peer reviewed.