Abstract
Background:
Many pharmacy services involve laboratory testing using point-of-care (POC) instruments. Our aim was to describe the implementation of quality control of the HbA1c POC instruments and investigate the performance in internal quality control (IQC) and external quality control (EQA) for HbA1c POC instruments in Norwegian community pharmacies.
Methods:
Two project pharmacists from each of 11 pharmacies participated in a training program covering capillary blood sampling, how to use the POC HbA1c instrument (DCA Vantage) and IQC and EQA. The pharmacies were enrolled in an EQA program for HbA1c, and their performance was compared with that of general practitioners’ (GP) offices.
Results:
Two of 89 (2.2%) IQC measurements were outside the acceptance interval. Seven out of 11 pharmacies sent in results in EQA in all the three surveys during six months. Two pharmacies did not analyze the control material in any of the surveys, one pharmacy analyzed the control material in one of the surveys, and one pharmacy analyzed two of the surveys. Of these pharmacies, 56-100% obtained “very good” evaluation for trueness and 71-100% obtained “very good” evaluation for precision. The corresponding numbers for GP offices were 75-87% for trueness and 84-94% for precision. No pharmacies obtained a “poor” evaluation.
Conclusions:
Norwegian community pharmacies can perform IQC and EQA on a HbA1c POC instrument, and the performance is comparable with that of GP offices. The compliance in the EQA surveys was modest, but the duration of the study and participation in the EQA program was probably too short to implement all the new procedures for all pharmacies.
Keywords: community pharmacy, diabetes, EQA, IQC, HbA1c, POC
Community pharmacists are easily accessible health care providers and have been characterized as an underutilized resource.1 Extended pharmacy services like medication counselling, disease management programs and screening are offered in community pharmacies in countries like the United Kingdom, New Zeeland, Australia, Canada, and Norway.2,3 Many of these services in pharmacies involve laboratory testing using point-of-care (POC) instruments. A prerequisite for a beneficial effect of these services is that they are performed in compliance with basic quality requirements for such instruments. That includes education and training of the users, and the need for, and adherence to, quality systems. Unreliable test results can lead to misdiagnosis, delayed treatment, and increased costs due to retesting. It is therefore important to ensure all results provided are both accurate and reliable for optimal patient care.
Previous studies from countries like Switzerland, Australia, and Spain have reported that it is feasible to implement services for diabetes risk assessment in a community pharmacy setting using HbA1c or glucose POC instruments.4-7 In addition, POC HbA1c instruments have been used to detect hyperglycemia and intervene to improve disease management when warranted in a community pharmacy setting.8,9 However, in these studies where pharmacies offered HbA1c or glucose measurements, quality control of the analysis was not reported.4-9 The need for quality control of POC testing is specified in guidelines,10,11 recommended in the literature,12,13 and implemented in several countries in settings like primary health care (PHC) and hospital laboratories.14-18 Pharmacies represent another setting compared with PHC and hospital laboratories. In Norway, while the pharmacy education training includes a great deal of laboratory work, they have little training in clinical chemistry like using POC instruments and quality control of the instruments. In Norway, 99% of the GP offices, all hospital laboratories and the majority of nursing homes participate in the Norwegian quality improvement of laboratory examinations (Noklus). In addition to external quality assessment (EQA) surveys for laboratory analysis, Noklus provides individual feedback, guidance in implementation of analysis including correct sampling and analysis, site visits, and laboratory and online courses for all participants in PHC.19 Noklus laboratory advisors also assist the participants in relation to performance, as well as interpretation and documentation of the internal quality control (IQC) results. The most important role of the laboratory advisors is to guide the users when deviant quality control results are found.
So far, pharmacies are not routinely offered enrolment in Noklus. However, in one previous study pharmacies were trained in how to perform IQC and participated in the EQA survey for glucose provided by Noklus.20 The quality of the performance of pharmacies was similar to that of GP offices.20
In an ongoing project, our aim is to investigate if Norwegian community pharmacies can contribute to identify persons with type 2 diabetes (T2D) using Finnish Diabetes Risk Score (FINDRISC)21 or QDiabetes®22 risk tests followed by HbA1c measurements using POC instruments for those at high risk. Our aim of this study was to describe the implementation of analytical quality control of the HbA1c POC instrument in community pharmacies, and their IQC and EQA performance for HbA1c POC instrument in this project.
Methods
In this longitudinal study, all 350 community pharmacies in a Norwegian pharmacy chain (Apotek 1) were invited to participate. A total of 19 pharmacies accepted, of which 11 were randomized to offer both a diabetes risk assessment and measurement of HbA1c and 8 offered only a diabetes risk assessment. The HbA1c POCT DCA Vantage Analyser (Siemens Healthcare Diagnostics, Erlangen, Germany) was selected for the study.23
Between 470 and 477 of the GP offices in Norway that are enrolled in Noklus used DCA Vantage Analyser during the project period, and results from EQA for these GP offices were included in the study.
Training of Pharmacists
Two project pharmacists from each of the 11 pharmacies offering measurement of HbA1c participated in a training program, and the pharmacies were enrolled in Noklus’ EQA program for HbA1c. At one pharmacy, one of the pharmacists left during the project period, and a new pharmacist was trained by the project leader. The pharmacists received the following training related to testing and analytical quality control:
- Completed two Noklus online courses covering capillary blood sampling and quality assurance:
- Capillary blood sampling: Preanalytical aspects, equipment to be used, theory, capillary blood sampling step by step, a film showing the whole procedure, ethics and control questions. The whole course takes about 25 minutes to complete.
- Quality assurance: Internal and external quality control, a film showing the whole procedure for internal quality control, differences between methods, documentation. The whole course takes about 25 minutes to complete.
Received written procedures describing how to perform HbA1c measurements.
- Attended a one-day course where they:
- Received oral and practical information about the HbA1c POCT DCA Vantage Analyser
- Had practical exercises using the HbA1c POCT DCA Vantage with guidance from a representative from the manufacturer
- Received oral information on how to perform internal and external quality control, and training with guidance from a representative from the manufacturer and a Noklus laboratory advisor
Received an on-site guidance by a Noklus laboratory advisor who ensured that they performed the capillary blood sampling and HbA1c measurement correctly.
As participants in Noklus, the pharmacies were offered guidance from a Noklus laboratory advisor by site visits, and the advisors were also available by telephone and email during working hours.
GP Offices
The GP offices have been enrolled in Noklus since 1993. They have a site visit from a Noklus laboratory advisor about every second year, and the employees at the GP offices involved in laboratory work are encouraged to take the Noklus online courses.
Internal and External Quality Controls for HbA1c
HbA1c IQCs in two levels from the producer of DCA Vantage Analyser were used, normal (target value 5.3% HbA1c (34 mmol/mol HbA1c)) and abnormal (target value 10.4% HbA1c (90 mmol/mol HbA1c)). The pharmacists were instructed to perform IQCs analysis each day a customer HbA1c measurement was performed and upon opening a new box of reagent. IQCs were evaluated against limits set by Noklus (acceptance interval = target value ± 10%).
Three Noklus EQA surveys for HbA1c were distributed during the study. The EQA control material for HbA1c from Noklus is pooled, fresh K2-EDTA blood in two levels from persons with and without diabetes (patient-like material). The target values are established by the European Reference Laboratory (ERL) for glycohemoglobin (the Netherlands). In the EQA survey for HbA1c the participants are evaluated for trueness and precision. Trueness is defined as the degree of closeness of the mean of the participants’ duplicate HbA1c measurements and the target value, and precision is given by the difference between the duplicate measurements of the controls. The Noklus criteria for assessing trueness and precision in the EQA survey for HbA1c are based on consensus of clinical outcome (Table 1). Only participants who returned results in both levels were included in the calculations.
Table 1.
Noklus Criteria for Assessing Trueness and Precision in the External Quality Assessment (EQA) Survey for HbA1c.
| Criterion | ||
|---|---|---|
| Trueness | ||
| Target interval | Defined as the target value ±0.1% HbA1c | |
| Very good | Mean value of duplicate measurements within ±2% of the target interval | |
| Acceptable | Mean value of duplicate measurements between the interval for “very good” and “poor” | |
| Poor | Mean value of duplicate measurements outside ±5.4% of the target interval. | |
| Precision | ||
| Very good | Difference between duplicate measurements ≤0.2 | |
| Acceptable | Difference between duplicate measurements between 0.3 and 0.4 | |
| Poor | Difference between duplicate measurements ≥0.5 | |
| Examples | ||
| Target value | 6.22% HbA1c | 44 mmol/mol |
| Target interval | 6.12-6.32% HbA1c | 43-46 mmol/mol |
| Very good trueness | 6.00-6.45% HbA1c | 42-47 mmol/mol |
| Acceptable trueness | 5.79-5.99, 6.46-6.66% HbA1c | 40-42, 47-49 mmol/mol |
| Poor trueness | ≤5.78, ≥6.67% HbA1c | <40, >49 mmol/mol |
The participants receive a feedback report with evaluation of their results with respect to trueness and precision and compared with other participants of the survey. Participants with “poor” results are contacted by a Noklus laboratory advisor to uncover the causes and correct any mistakes that may have been made.
After the study period, surveys were sent to the Noklus laboratory advisors and the pharmacists. Noklus laboratory advisor were asked about the cooperation with the pharmacies, if they had any contact after the initial visit and the routines for follow-up when participants do not answer the EQA survey. Questions to the pharmacists included if they had taken the online courses and whether they found them useful. In addition, oral information was collected about their experience with the project, particularly in regard to measurements and quality control of HbA1c.
Statistics
Statistical analysis was performed using R version 3.4.1.24 The lme4 package using the lmer function for a linear mixed-effects model25 was used for calculation of coefficient of variation (CV) with 95% confidence interval (CI) for IQC from all pharmacies (CV%Between-pharmacies) and operators (pharmacists) (CV%Between-operator). The lmer function for a random nested model25 was used for calculation of the imprecision with 95% CI for the EQA control within (CV%Within) and between (CV%Between) pharmacies or GP offices. Excel version 15.38 was used for descriptive statistical analysis and figure.
HbA1c is reported in both SI (mmol/mol) and % units using the master equation: NGSP (HbA1c, %) = [0.09148 * IFCC (HbA1c, mmol/mol)] + 2.152. The criteria for trueness and precision in Table 1 are reported only in percentage HbA1c units.
Results
Three project pharmacists did not attend the one-day course, but they received personal instructions and training by the project leader and Noklus laboratory advisor. In total 75 HbA1c measurements were performed at nine of the pharmacies during the study. Two pharmacies did not perform any HbA1c measurements.
Results From IQC
In total 89 IQC measurements were reported from eight pharmacies (Table 2). Results from one pharmacy are missing, and at two pharmacies no IQC measurements was performed because they did not perform any HbA1c measurements on customers. In total, there were 15 operators, and they each reported between one and 14 of the 89 IQC results (Figure 1). The mean HbA1c was within the acceptance interval in both levels, while two pharmacies reported one measurement each outside the acceptance interval (Table 2, Figure 2). The CV%Between-pharmacies (95% CI) was 1.9% (0.83, 3.9) and 5.0% (3.0, 9.8) in the normal and abnormal level, respectively. The corresponding numbers for operators (CV%Between-operators) were 0.91 (0, 2.5) and 5.5 (2.1, 13.0). A total of 40 customer HbA1c measurements were performed on days where the pharmacy performed IQC measurements, while 31 customer HbA1c measurements were performed without IQC the same day. And 11 customer HbA1c measurements were performed without IQC the same week.
Table 2.
Results From Internal Quality Control (IQC) From the Pharmacies.
| Control material |
||
|---|---|---|
| Level 1 (normal) (n = 58) | Level 2 (abnormal) (n = 31) | |
| Number of IQC measurements in each pharmacy (mean ± SD) | 7.3 ± 3.9 | 5.0 ± 2.4 |
| Target (acceptance intervala), (HbA1c, %) | 5.3 (4.8-5.8) | 10.4 (9.4-11.4) |
| Target (acceptance intervala), (HbA1c, mmol/mol) | 34 (29-40) | 90 (79-101) |
| HbA1c (%), (mean (95% CI)) | 5.2 (5.1, 5.3) | 10.5 (10.0, 11.0) |
| HbA1c (mmol/mol), (mean (95% CI)) | 33 (32, 34) | 86 (79, 93) |
| Total number of IQC measurements outside acceptance interval | 1 | 1 |
| CV%Between-pharmacies (95% CI) | 1.9 (0.83, 3.9) | 5.0 (3.0, 9.8) |
Acceptance interval = target value ±10%.
Figure 1.
Frequency histogram showing the number of internal quality controls (IQCs) conducted by each of the 15 different operators (pharmacists) at eight different pharmacies. Mean number of measurements by each operator (± SD): 4.9 (± 3.3) and 2.1 ± (2.2) in level 1 and level 2, respectively.
Figure 2.
Frequency histograms of the internal quality (IQ) data for level 1 (normal) and level 2 (abnormal) obtained by the 15 different operators at eight different pharmacies. Acceptance interval in level 1: 4.8-5.8% HbA1c (29-40 mmol/mol HbA1c); acceptance interval in level 2: 9.4-11.4% HbA1c (79-101 mmol/mol HbA1c).
Results from External Quality Assessment (EQA)
Seven pharmacies participated in all the three Noklus EQA HbA1c surveys in the period from October 2016 to April 2017. One pharmacy analyzed the control material in two of these surveys, and one pharmacy analyzed the control material in one of these surveys. Two pharmacies did not analyze the control material in any of the surveys (Tables 3 and 4). In total, 441, 424, and 402 GP offices using DCA Vantage reported results in the same three Noklus EQA HbA1c surveys, respectively. Thus, between 19% (n = 2) and 36% (n = 4) of the pharmacies and between 5.5% and 11.5% of the GP offices that received EQA control material did not return results and are not included in the calculations of trueness and precision.
Table 3.
Results From the Noklus External Quality Assessment (EQA) Surveys for HbA1c Showing the Percentage and Number of Pharmacies and Percentage of General Practitioner (GP) Offices Obtaining “Very Good,” “Acceptable,” and “Poor” Evaluations for Trueness.
| Level | Target value (target
interval), HbA1c |
Truenessa |
|||||||
|---|---|---|---|---|---|---|---|---|---|
| Very good |
Acceptable |
Poor |
|||||||
| % | mmol/mol | Pharmacies, % (n) | GP offices, % | Pharmacies, % (n) | GP offices, % | Pharmacies, % (n) | GP offices, % | ||
| Survey 1 | 1 | 5.20 (5.10-5.30) | 33 (32-34) | 87.5 (7) | 84.6 | 12.5 (1) | 12.5 | 0.0 (0) | 0.91 |
| 2 | 7.09 (6.99-7.19) | 54 (53-55) | 100 (8) | 87.3 | 0.0 (0) | 11.8 | 0.0 (0) | 0.23 | |
| Survey 2 | 1 | 5.96 (5.86-6.06) | 42 (41-43) | 55.6 (5) | 80.4 | 44.4 (4) | 18.2 | 0.0 (0) | 0.71 |
| 2 | 7.07 (6.97-7.17) | 54 (53-55) | 100 (8) | 85.6 | 0.0 (0) | 13.2 | 0.0 (0) | 0.24 | |
| Survey 3 | 1 | 5.52 (5.42-5.62) | 37 (36-38) | 100 (7) | 87.1 | 0.0 (0) | 11.7 | 0.0 (0) | 0.75 |
| 2 | 6.22 (6.12-6.32) | 44 (43-46) | 100 (7) | 75.4 | 0.0 (0) | 22.6 | 0.0 (0) | 1.24 | |
Calculations are based on participants who returned results in both levels in the EQA surveys. Between 19% (n = 2) and 36% (n = 4) of the pharmacies and between 5.5% and 11.5% of the GP offices did not return results.
“Very good”: deviation of results from the target interval within ±2% of the target interval. “Poor”: deviation of results from the target interval outside ±5.4% of the target interval. “Acceptable”: deviation of results from the target interval between the interval for “very good” and “poor.” The numbers of GP offices were 441, 424, and 402 in surveys 1, 2, and 3, respectively. Participants with “poor” evaluation for precision (difference between duplicate measurement ≥ 0.5) (Table 4) are not evaluated for trueness. Thus, the sums of the percentage values across the table are less than 100% for the GP offices.
Table 4.
Results From the Noklus External Quality Assessment (EQA) Surveys for HbA1c Showing the Percentage and Number of Pharmacies and Percentage of General Practitioner (GP) Offices Obtaining “Very Good,” “Acceptable,” and “Poor” Evaluations for Precision.
| Level | Precisiona |
||||||
|---|---|---|---|---|---|---|---|
| Very good |
Acceptable |
Poor |
|||||
| Pharmacies, % (n) | GP offices, % | Pharmacies, % (n) | GP offices, % | Pharmacies, % (n) | GP offices, % | ||
| Survey 1 | 1 | 100 (8) | 93.4 | 0.0 (0) | 4.54 | 0.0 (0) | 2.04 |
| 2 | 87.5 (7) | 84.1 | 12.5 (1) | 15.2 | 0.0 (0) | 0.68 | |
| Survey 2 | 1 | 100 (7) | 92.2 | 0.0 (0) | 7.1 | 0.0 (0) | 0.71 |
| 2 | 100 (7) | 85.9 | 0.0 (0) | 13.2 | 0.0 (0) | 0.94 | |
| Survey 3 | 1 | 71.4 (5) | 93.5 | 28.6 (2) | 6.0 | 0.0 (0) | 0.50 |
| 2 | 100 (7) | 92.0 | 0.0 (0) | 7.2 | 0.0 (0) | 0.75 | |
Calculations are based on participants who returned results in both levels in the EQA surveys. Between 27% (n = 3) and 36% (n = 4) of the pharmacies and between 7.1% and 14.0% of the GP offices did not return results.
“Very good”: difference between duplicate measurement ≤0.2. “Acceptable”: difference between duplicate measurement between 0.3 and 0.4. “Poor”: difference between duplicate measurement ≥0.5. The numbers of GP offices were 441, 424, and 402 in surveys 1, 2, and 3, respectively.
Between 56% and 100% of the pharmacies obtained a “very good” evaluation for trueness (Table 3) and between 71% and 100% obtained a “very good” evaluation for precision (Table 4). The corresponding numbers for GP offices were 75-87% for trueness (Table 3) and 84-94% for precision (Table 4). No pharmacies obtained a “poor” evaluation, while 0.23-1.24% and 0.50-2.04% of the GP offices obtained a “poor” evaluation for trueness (Table 3) and precision (Table 4), respectively. For the pharmacies, the deviation from target (trueness) tended to be lowest in survey 3, and all participating pharmacies obtained “very good” evaluation for trueness (Table 3). However, difference between duplicate measurements (precision) was lowest in survey 2 where 100% obtained “very good” evaluation for precision (Table 4).
Differences between pharmacies and GP offices in CV%Within and CV%Between were found for sample 1 in survey 1, CV%Within (95% CI) pharmacies: 1.06 (0.69, 1.90) vs GP offices: 2.04 (1.91, 2.18) and sample 2 in survey 2, CV%Between (95% CI) pharmacies: 0.75 (0.48, 1.40) vs GP offices: 1.75 (1.64, 1.88) (Table 5). However, the results must be interpreted with caution since the number of pharmacies is very low (n = 7-9) and the 95% CI is wide compared to the GP offices (n = 402-441) (Table 5).
Table 5.
Results From the Three Noklus External Quality Assessment (EQA) Surveys for HbA1c for the Pharmacies and for the General Practitioner (GP) Offices.
| Sample 1 |
Sample 2 |
||||
|---|---|---|---|---|---|
| Pharmacies | GP offices | Pharmacies | GP offices | ||
| Survey 1 | Number of participants | 8 | 441 | 8 | 441 |
| Target value (target interval), HbA1c (%) | 5.20 (5.10-5.30) | 7.09 (6.99-7.19) | |||
| Target value (target interval), HbA1c (mmol/mol) | 33 (32-34) | 54 (53-55) | |||
| Mean (95% CI), % HbA1c | 5.31 (5.21, 5.40) | 5.31 (5.30, 5.33) | 7.05 (7.02, 7.17) | 7.16 (7.15, 7.18) | |
| Mean (95% CI), mmol/mol | 35 (33, 36) | 35 (34, 35) | 54 (53, 55) | 55 (55, 55) | |
| CV% within (95% CI) | 1.06 (0.69, 1.90) | 2.04 (1.91, 2.18) | 1.70 (1.11, 2.61) | 2.14 (2.00, 2.28) | |
| CV% between (95% CI) | 2.19 (1.26, 4.08) | 1.70 (1.49, 1.92) | 0.50 (0.0, 2.03) | 1.50 (1.27, 1.72) | |
| Survey 2 | Number of participants | 7 | 424 | 7 | 424 |
| Target value (target interval), HbA1c (%) | 5.96 (5.86-6.06) | 7.07 (6.97-7.17) | |||
| Target value (target interval), HbA1c (mmol/mol) | 42 (41-43) | 54 (53-55) | |||
| Mean (95% CI), % HbA1c | 6.08 (5.97, 6.20) | 6.07 (6.06, 6.08) | 7.05 (6.95, 7.15) | 7.17 (7.15, 7.18) | |
| Mean (95% CI), mmol/mol | 43 (42, 44) | 43 (43, 43) | 54 (52, 55) | 55 (55, 55) | |
| CV% within (95% CI) | 2.40 (1.38, 4.33) | 1.68 (1.50, 1.87) | 1.73 (1.04, 3.22) | 1.56 (1.38, 1.75) | |
| CV% between (95% CI) | 1.17 (0.75, 2.17) | 1.73 (1.62, 1.89) | 0.75 (0.48, 1.40) | 1.75 (1.64, 1.88) | |
| Survey 3 | Number of participants | 7 | 402 | 7 | 402 |
| Target value (target interval), HbA1c (%) | 5.52 (5.42-5.62) | 6.22 (6.12-6.32) | |||
| Target value (target interval), HbA1c (mmol/mol) | 37 (36-38) | 44 (43-46) | |||
| Mean (95% CI), % HbA1c | 5.50 (5.42, 5.58) | 5.57 (5.56, 5.58) | 6.26 (6.18, 6.33) | 6.35 (6.33, 6.36) | |
| Mean (95% CI), mmol/mol | 37 (36, 37) | 37 (37, 37) | 45 (44, 46) | 46 (46, 46) | |
| CV% within (95% CI) | 0.00 (0.00, 2.67) | 1.83 (1.64, 2.03) | 0.99 (0.00, 2.42) | 2.13 (1.94, 2.33) | |
| CV% between (95% CI) | 2.48 (1.64, 3.78) | 1.76 (1.64, 1.89) | 1.36 (0.87, 2.42) | 1.59 (1.49, 1.71) | |
The measurement of the control material at the GP offices were mainly performed by medical secretaries (76-78%), nurses (10-11%) and biomedical laboratory scientists (10%).
Follow-Up by Noklus Laboratory Advisors
All Noklus laboratory advisors answered the survey. The contact with the pharmacies after the initial visit varied between the sites due to heavy work load, sick leave, and different routines. One Noklus laboratory advisor sent an email to the pharmacy due to high IQC value, but the pharmacy did not respond. The others had one or two visits, one or two contacts by phone or email correspondence. One Noklus laboratory advisor had email and telephone contact and visited the pharmacy do discuss routines for IQC, storage and expiration of the reagents. One Noklus laboratory advisor did not consider the pharmacies as fully members of Noklus since they were a part of a project. Normally the Noklus laboratory advisors follow up with participants when they do not submit results from EQA over time. None of the pharmacies were contacted due to missing response to EQA survey.
Noklus Online Courses
In all, 14 of the 22 pharmacists answered the survey. At three pharmacies both pharmacists answered the survey, at eight pharmacies one pharmacists answered the survey, and two pharmacies did not answer. All these pharmacists took the two Noklus online courses, and 11 found the courses useful, while three found them to be too detailed or did not think they were relevant.
Discussion
This study shows that it is possible to implement quality control of HbA1c POC instrument in pharmacies. For the IQC the between-laboratory CV was 1.9% and 5.0% in normal and abnormal HbA1c level, respectively (Table 2). Thus, the pharmacies fulfilled the recommendation of between-laboratory CV <3% for a single method26 in the normal level. In addition, only 2% of the IQC measurement were outside the acceptance interval (Table 2, Figure 2). In the HbA1c EQA surveys no pharmacies, and only a small percentage (0.23-2.0%) of the GP offices obtained “poor” evaluation for trueness and precision (Tables 3 and 4). The percentages of pharmacies obtaining “very good” and “acceptable” evaluation for trueness and imprecision were comparable with GP offices (Tables 3 and 4). In addition, the analytical variation and between variation was comparable with results obtained by GP offices (Table 5).
Strengths of this study are that the pharmacies were enrolled in an EQA organization with 25 years of experience,19 and the Noklus laboratory advisors were available for any questions the pharmacist may have regarding IQC and EQA. The pharmacies were enrolled in Noklus for a period of six months and participated in three EQA surveys, so their performance was monitored over time and not just at one time point. It has been showed that participation in EQA programs improve analytical quality over time.27,28 We did not observe improvement in the pharmacies’ performance during the length of this study, but six months is probably too short to show improvement in analytical quality, especially when they started at a relatively high level. This might be due to that the training program was satisfactory, and that the HbA1c POC instrument that was chosen is reported to be user-friendly and fulfil the analytical quality specifications.23 The number of GP offices was high, and both the pharmacies and GP offices represents various locations throughout the country. However, the pharmacies were self-recruited so they might not be representative of all pharmacies.
Limitations of this study are that the number of pharmacies is low and that only nine reported EQA results. Thus, the analytical variation and variation between the pharmacies are only indications of their performance compared to results obtained by more than 400 GP offices.
IQC was performed on the same day as a customer was tested for only half of the customers HbA1c measurements. However, only 15% of the customers HbA1c measurements were performed without IQC same week. Noklus laboratory advisors instruct other Noklus participants to perform IQC every week. In this study, it was decided to perform IQC more often to secure correct HbA1c measurement. Thus, there might have been unclear instructions for how often to perform IQC at the course day and from the Noklus laboratory advisors.
The two outlying IQC measurements were obtained by two different operators with the same level of training and experience at two different pharmacies, and were the second and third IQC measurement, respectively, for each of the two operators. Thus, there is no relationship between outliers and who performed the analysis. We do not have complete data about each reagent lot that were used, and effects of different lots can therefore not be investigated. Nevertheless, two different lots were used for the two outlying IQC results.
A tight schedule at the one-day course may have resulted in too little time to stress the importance of IQC and EQA. Other participants in Noklus have a 3 to 4-hour course only about quality control and blood sampling. Still, the pharmacy participation is quite similar to Noklus’s experiences with other participants: the compliance in EQA surveys and reporting IQC results for GP offices and nursing homes are also modest, especially the first year after enrolment in Noklus. In a previous study in a pharmacy setting all participants reported results in the EQA surveys.20 However, in that study the implementation of quality control was the main focus, while in this study IQC and EQA was one of several new procedures.
When participants obtain “poor” evaluations in the Noklus EQA surveys, they are contacted by a Noklus laboratory advisor, which are experienced biomedical laboratory scientists. Together the participant and the advisor try to uncover the causes for the deviant result and correct any mistakes that might have been done. Furthermore, Noklus offers guidance for all participants, site visits, laboratory and E-learning courses, in addition to producing guidelines and procedures for all laboratory work performed in PHC. If the result of a IQC sample is outside the set limits, the patient samples should not be reported until the reason for the error is determined and corrected. The participants are encouraged to contact Noklus laboratory advisors if they have questions. Noklus offers different forms, including electronic forms that participants can use for documentation of IQC results. These are all actions to prevent future HbA1c “errors.”
While one study in a Norwegian setting showed that pharmacists can obtain EQA results similar to that of GP offices for glucose,20 a study from Italy found that the analytical performance of POCT for glucose, cholesterol and triglycerides in three different pharmacies was relatively heterogeneous and two to five times higher compared with laboratory technicians.12 A study from 2008 found that only 37% of the Norwegian community pharmacies that offer to check the patients’ blood glucose instrument reported to perform analytical quality control of the analysis.29 In Scotland, only 6.5% pharmacist reported that they have a quality assurance in place for different services, and analytical quality control was not included.30 In addition, in previous studies where measurement of glucose or HbA1c is performed in pharmacies, analytical quality control is not reported.4-9 In this study, some of the pharmacists did not understand the importance of quality control, and prioritized the recruitment of customers and the conduct of the service before the quality control.
This shows the need to focus on the importance of analytical quality control in pharmacies where POC testing is offered. There is no regulation that describes education or training or minimum clinical knowledge. To implement quality control, it is essential that the pharmacists receive education in how to perform the analytical quality control. Furthermore, it takes time to get good routines for laboratory work and to become familiar with analysis and IQC and EQA. Thus, follow up by the EQA organization (Noklus laboratory advisor) is important, especially the first year after enrolment. In this study, unclear instructions to the pharmacist and Noklus laboratory advisors may have been the reason that the IQC not was performed as advised, and that the pharmacists did not always contacted the Noklus laboratory adviser if they had problems.
Conclusions
The pharmacies show the ability to fulfil the recommendations from international organizations and Noklus for internal and external analytical quality for POC HbA1c measurement. The performance in the EQA surveys was comparable to results obtained by GP offices with many years of experience. IQC and EQA should be mandatory for all pharmacies offering services involving POCT to ensure accurate and reliable results for optimal patient care.
Acknowledgments
We thank Thomas Helge Røraas for valuable statistical help.
Footnotes
Abbreviations: CI, confidence interval; EQA, external quality assessment; ERL, European Reference Laboratory; FINDRISC, Finnish Diabetes Risk Score; GP, general practitioner; HbA1c, hemoglobin A1c; IFCC, International Federation of Clinical Chemistry and Laboratory Medicine; IQC, internal quality control; Noklus, Norwegian Quality Improvement of Laboratory Examinations; PHC, primary health care; POC, point of care; T2D, type 2 diabetes.
Declaration of Conflicting Interests: The author(s) declared no potential conflicts of interest with respect to the research, authorship, and/or publication of this article.
Funding: The author(s) disclosed receipt of the following financial support for the research, authorship, and/or publication of this article: The study was financed by The Norwegian Community Pharmacy Foundation (https://www.proff.no/selskap/stiftelsen-til-fremme-av-norskapotekfarmasi/oslo/fondlegat/IG9KD0M10NC/) and the University of Bergen (http://www.uib.no). The purpose of the Norwegian Community Pharmacy Foundation is to help to promote the development of Norwegian pharmacy. Siemens (https://www.siemens.com/no/no/home.html) lent the DCA
Vantage Analyzer instruments to the pharmacies and provided training free of charge. The participating pharmacy chain, Apotek 1, covered expenses of the course day and paid the membership fee in Noklus. The Norwegian Community Pharmacy Foundation had no influence on how the study was conducted, the analysis or the results. Apotek 1, University of Bergen, Siemens and The Norwegian Community Pharmacy Foundation had no influence on the analysis or the results, but Apotek 1 have been given valuable input to the protocol and procedures.
ORCID iD: Aslaug Johanne Risøy 
https://orcid.org/0000-0002-0311-1174
References
- 1. Murray R. Community pharmacy clinical services review. NHS commissioning. 2017. Available at: https://www.england.nhs.uk/commissioning/wp-content/uploads/sites/12/2016/12/community-pharm-clncl-serv-rev.pdf. Accessed March 19, 2018.
- 2. Wright D. A rapid review of evidence regarding clinical services commissioned from community pharmacies. 2016. Available at: https://www.england.nhs.uk/commissioning/wp-content/uploads/sites/12/2016/12/rapid-evdnc-rev-dec-16.pdf. Accessed March 19, 2018.
- 3. Hovland R. Farmasøyttjenester. Apokus. 2014. Available at: http://www.apotek.no/Files/Filer_2014/Rapporter/2014-05-06%20Rapport%20farmasøyttjenester.pdf. Accessed March 19, 2018.
- 4. Hersberger KE, Botomino A, Mancini M, Bruppacher R. Sequential screening for diabetes—evaluation of a campaign in Swiss community pharmacies. Pharm World Sci. 2006;28(3):171-179. [DOI] [PubMed] [Google Scholar]
- 5. Krass I, Mitchell B, Clarke P, et al. Pharmacy diabetes care program: analysis of two screening methods for undiagnosed type 2 diabetes in Australian community pharmacy. Diabetes Res Clin Pract. 2007;75(3):339-347. [DOI] [PubMed] [Google Scholar]
- 6. Fikri-Benbrahim N, Martínez-Martínez F, Saéz-Benito L, et al. Assessment of a screening protocol for type 2 diabetes in community pharmacy. The DiabNow Study. Diabetes Res Clin Pract. 2015;108(3):e49-e52. [DOI] [PubMed] [Google Scholar]
- 7. Fornos-Pérez JA, Andrés-Rodríguez NF, Andrés-Iglesias JC, et al. Detection of people at risk of diabetes in community pharmacies of Pontevedra (Spain) (DEDIPO). Endocrinol Nutr. 2016;63(8):387-396. [DOI] [PubMed] [Google Scholar]
- 8. Papastergiou J, Rajan A, Diamantouros A, Zervas J, Chow J, Tolios P. HbA1c testing in the community pharmacy: a new strategy to improve care for patients with diabetes. Can Pharm J (Ott). 2012;145(4):165-167. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 9. Papastergiou J, Folkins C, Li W. Community pharmacy-based A1c screening: a Canadian model for diabetes care. Int J Pharm Pract. 2016;24(3):189-195. [DOI] [PubMed] [Google Scholar]
- 10. International Organization for Standardization. ISO 22870:2005 Standard. Point-of-care testing (POCT)—Requirements for quality and competence. Geneva: ISO; 2006. [Google Scholar]
- 11. Clinical and Laboratory Standards Institute. Quality management: approaches to reducing errors at the point of care; approved guideline. Report POCT07-A. Waynes, PA: Clinical and Laboratory Standards Institute; 2010. [Google Scholar]
- 12. Lippi G, Plebani M, Favaloro EJ, Trenti T. Laboratory testing in pharmacies. Clin Chem Lab Med. 2010;48(7):943-953. [DOI] [PubMed] [Google Scholar]
- 13. Gronowski AM, Adams A, Ball C, Gaydos CA, Klepser M. Pharmacists in the laboratory space: friends or foes? Clin Chem. 2016;62(5):679-683. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 14. Shephard MDS, Gill JP. The analytical quality of point-of-care testing in the “QAAMS” model for diabetes management in Australian aboriginal medical services. Clin Biochem Rev. 2006;27(4):185-190. [PMC free article] [PubMed] [Google Scholar]
- 15. Wood WG. Problems and practical solutions in the external quality control of point of care devices with respect to the measurement of blood glucose. J Diabetes Sci Technol. 2007;1(2):158-163. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 16. Kitchen DP, Kitchen S, Jennings I, Woods T, Walker I. Quality assessment in point-of-care coagulation testing. Semin Thromb Hemost. 2008;34(7):647-653. [DOI] [PubMed] [Google Scholar]
- 17. Thomas MA, Sall S, Perkins M, Blount N, Ducroq DH. An external quality assessment scheme for point of care testing in high street pharmacies—the UK model: W-a233. Clin Chem Lab Med. 2009;47(Suppl):S324. [Google Scholar]
- 18. Stavelin A, Meijer P, Kitchen D, Sandberg S. External quality assessment of point-of-care International Normalized Ratio (INR) testing in Europe. Clin Chem Lab Med. 2012;50(1):81-88. [DOI] [PubMed] [Google Scholar]
- 19. Norwegian Quality Improvement of Laboratory Examinations. Available at: http://www.noklus.no/en/Home.aspx. Accessed October 13, 2017.
- 20. Kjome RLS, Nerhus K, Sandberg S. Implementation of a method for glucose measurements in community pharmacies. Int J Pharm Pract. 2010;18(1):13-19. [DOI] [PubMed] [Google Scholar]
- 21. Lindström J, Tuomilehto J. The diabetes risk score: a practical tool to predict type 2 diabetes risk. Diabetes Care. 2003;26(3):725-731. [DOI] [PubMed] [Google Scholar]
- 22. Hippisley-Cox J, Coupland C, Robson J, Sheikh A, Brindle P. Predicting risk of type 2 diabetes in England and Wales: prospective derivation and validation of QDScore. BMJ. 2009;338:b880. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 23. Sølvik UØ, Roraas T, Christensen NG, Sandberg S. Diagnosing diabetes mellitus: performance of hemoglobin A1c point-of-care instruments in general practice offices. Clin Chem. 2013;59(12):1790-1801. [DOI] [PubMed] [Google Scholar]
- 24. R Core Team. R: a language and environment for statistical computing. Vienna, Austria: R Foundation for Statistical Computing; nd; Available at: https://www.R-project.org. Accessed October 13, 2017. [Google Scholar]
- 25. Bates D, Maechler M, Bolker B, et al. lme4: linear mixed-effects models using “Eigen” and S4. Available at: https://cran.r-project.org/web/packages/lme4/index.html. Accessed October 13, 2017.
- 26. Sacks DB, Arnold M, Bakris GL, et al. Guidelines and recommendations for laboratory analysis in the diagnosis and management of diabetes mellitus. Diabetes Care. 2011;34(6):e61-e99. [DOI] [PMC free article] [PubMed] [Google Scholar]
- 27. Kristensen GBB, Nerhus K, Thue G, Sandberg S. Results and feasibility of an external quality assessment scheme for self-monitoring of blood glucose. Clin Chem. 2006;52(7):1311-1317. [DOI] [PubMed] [Google Scholar]
- 28. Bukve T, Stavelin A, Sandberg S. Effect of participating in a quality improvement system over time for point-of-care C-reactive protein, glucose, and hemoglobin testing. Clin Chem. 2016;62(11):1474-1481. [DOI] [PubMed] [Google Scholar]
- 29. Kjome RLS, Sandberg S, Granås AG. Diabetes care in Norwegian pharmacies: a descriptive study. Pharm World Sci. 2008;30(2):191-198. [DOI] [PubMed] [Google Scholar]
- 30. Douglas E, Power A, Hudson S. Pharmaceutical care of the patient with diabetes mellitus: pharmacists’ priorities for services and educational needs in Scotland. Int J Pharm Pract. 2007;15(1):47-52. [Google Scholar]