Sage Journals: Discover world-class research

Abstract

Background

The Dutch guideline algorithm for the analysis of anaemia in patients of general practitioners (GPs) was programmed in a Clinical Decision Support system (CDS-anaemia) to support the process of diagnosing the cause of anaemia in the laboratory. This study investigates the diagnostic yield of the automated anaemia algorithm compared to that of the manual work up by the GP.

Methods

This retrospective population-based study consisted of 2697 people ≥18 years. Anaemia was defined according to the Dutch College of General Practitioners (DCGP) guideline. Causes of anaemia were based on the DCGP guidelines with the corresponding blood tests. The number of blood tests and causes of anaemia were measured in two separate periods in both the (CDS-anaemia) pilot group and a control group in which routine care was provided.

Results

Patients from GPs supported by CDS-anaemia had higher chances of having more anaemia-related blood tests being performed. This resulted in finding significantly more causes of anaemia in the pilot group compared to the control group with respect to iron deficiency (resp. 31.3% vs 14.5%), possible iron deficiency (resp. 11.4% vs 2.8%), iron deficiency in acute phase (2.6% vs 0.5%), chronic disease/infection/inflammation (23.5% vs 1.9%), vitamin B12 deficiency (4.5% vs 1.9%), possible vitamin B12 deficiency (16.8% vs 8.7%), folate deficiency (3.3% vs 0.9%) and possible bone marrow disorder (3.8% vs 0.0%); p < 0.05.

Conclusions

This study suggests that an automated-algorithm support can effectively aid in the diagnostic work-up of anaemia in primary care to find more causes of anaemia.

Keywords

Anaemia effectiveness automated algorithm laboratory work-up general practitioners

Introduction

Anaemia is a highly prevalent disorder worldwide affecting 24.8% of the population.¹ Children (up to 40%), women (29% non-pregnant and 38% during pregnancy) and older individuals (ranging from 20 to 37%) are most often affected.^1–4 Anaemia is not a diagnosis but often a presentation of an underlying condition.^5,6 It is therefore relevant to diagnose the cause of anaemia to treat the patient as adequately as possible. Remarkably, this aspect has not been studied well, although some reports state that the cause of anaemia is often underdiagnosed.^7,8

General practitioners (GPs) need laboratory tests alongside anamnesis and physical examination in order to infer accurate conclusions about the cause of anaemia. The Dutch College of General Practitioners (DCGP) in the Netherlands provides a guideline for GPs to support medical decision making for anaemia.⁹ This guideline provides an algorithm which includes laboratory tests that need to be chosen under different circumstances. These tests consist of haemoglobin (Hb), mean corpuscular volume (MCV), ferritin, vitamin B12, folate, lactate dehydrogenase (LDH), leukocytes, thrombocytes, erythrocytes, reticulocytes, iron (Fe), transferrin, creatinine/estimated glomerular filtration rate (eGFR) and erythrocyte sedimentation rate (ESR).⁹

Despite the existing guidelines, a study among 1342 anaemic patients from GPs has shown that in 86% of the patients one or more laboratory tests were missing for explaining the cause of anaemia.⁷ We recently also targeted cause finding in anaemia by GPs. This study showed that a laboratory diagnosis is found only in 22,5% of the male patients aged 18–44, while a diagnosis is most often found in older females (49,1% of anaemic patients).⁸ Since anaemia is much more prevalent in older individuals, finding anaemia in young males is especially abnormal.³ It is important that laboratory tests are requested at the right time, that is, pursuing cause finding of anaemia only when anaemia is present. Otherwise overutilization of diagnostic blood tests might occur, but also underutilization in anaemic patients with the risk of incomplete or missing diagnosis and potentially leading to increased healthcare costs.^8–11

General practitioners may therefore benefit from added guidance from laboratory diagnostics to reveal the cause of anaemia.^10,12 To overcome missing causes of anaemia and supporting correct testing strategy, an automated algorithm was developed by Atalmedial in 2020 that supports the GP in finding the cause of anaemia according to the DCGP guidelines. The subsequent steps of the DCGP anaemia guideline were programmed in a Clinical Decision Support (CDS) software further called CDS-anaemia. This is an automated process, which is set in motion after blood samples are analysed for the presence of anaemia. Clinical decision support-anaemia automatically adds extra tests in a stepwise manner when anaemia is found. However, the effect of CDS-anaemia in supporting the GP with the diagnostic process and the cause finding of anaemia in primary care is unknown. Therefore, in the current study we investigated to which extent CDS-anaemia is effective in finding the possible cause(s) of anaemia in undiagnosed patients from GPs, compared to GPs who were not supported by CDS-anaemia (i.e. manual diagnostic work-up).

Methods

Study design

The data used for this analysis were obtained from the Atalmedial Medical Diagnostics laboratory request and results database. Data consisted of primary care patients, including sex, age and dates of measurement. Data also included the relevant blood parameters: Hb, erythrocytes, MCV, red cell distribution width (RDW), ferritin, iron, transferrin, vitamin B12, folate, LDH, reticulocytes, leukocytes, thrombocytes, creatinine, ESR and C-reactive protein (CRP). The GPs were located in the West of the Netherlands and included in total 53 GPs.

Pilot and control group

This study used two groups in order to establish the effectiveness of the intervention: a pilot group and a control group. The pilot group consisted of patients from GPs who were supported by Atalmedial Medical Diagnostics by providing automated-anaemia algorithm support when doing their blood diagnostics. The automated anaemia algorithm is performed by a software tool called AlinIQ (Abbott Diagnostics), hence referred to as CDS-anaemia. A distinction was made between data that originated from before the use and during the use of CDS-anaemia. Data derived from individuals during CDS-anaemia implementation, effective from 11–11–2020 till 1–3–2022 (second time period), was used to compare with the data from individuals pre-CDS-anaemia implementation from 11–11–2018 till 1–3–2020 (first time period). This resulted in a 15 months’ time period for both time frames. In addition, a control group was added, consisting of patients from GPs who did not receive CDS-anaemia assistance in their blood diagnostics, that is, did their work-up manually. This control group was added to control for the effects of the COVID-19 pandemic and the possible change in requesting behaviour by the GP over time. The same timeline was applied to each group to control for possible seasonal effects that could potentially influence results.

Patient characteristics

Four age groups were defined: 18–44, 45–64, 65–79 and >79 years. In order to measure effectiveness of the intervention in undiagnosed patients, it was assumed that only newly admitted patients within the Atalmedial system should be included. This was established by including patients in the time window as described above, who had Hb and indices (MCV, RDW and erythrocytes) measurements, but without an Hb measurement 6 months prior to the study Hb. Patients were excluded if they were <18 years. Since the manual follow up of anaemia by a GP according to the DCGP anaemia guideline will take time, we also included blood test results in the analysis that were requested within 6 weeks of the study Hb (second request).

Definitions and anaemia algorithm

Anaemia and the anaemia algorithm were defined according to the DCGP criteria.⁹ ESR in the DCGP guideline was replaced by CRP, as the test specifications favour use of this parameter in the algorithm. The DCGP criteria were programmed in CDS. The CDS algorithm is implemented at the laboratory site. Clinical decision support-anaemia was only activated for patients from GPs that participated in the pilot group. Clinical decision support-anaemia became activated in this group when Hb and indices were requested. Subsequently, CDS-anaemia added ferritin tests in all patients with anaemia. Depending on the results of ferritin and MCV, RDW and number of erythrocytes, a conclusion was generated or subsequent tests were added. The CDS-anaemia workflow is depicted in Figure 1. The combination of tests added is based upon the DCGP guideline anaemia.⁹ The results were approved by a specialist laboratory medicine, tests results and the conclusion were reported to the GP. The causes of anaemia were defined according to the conditions specified in Table 1.

Figure 1.

Flowchart clinical decision support-anaemia algorithm.

Table 1.

Definition of causes of anaemia as defined in CDS-anaemia.

Possible cause of anaemia	Condition male	Condition female
Iron deficiency	Ferritin <25 μg/L	Ferritin <10 μg/L
Likely iron deficiency*	25 ≤ Ferritin ≤30 μg/L	10 ≤ Ferritin ≤ 25 μg/L
Iron deficiency in acute phase	30 < Ferritin <100 μg/L; Iron <14 μmol/; CRP >5 mg/L	25 < Ferritin <100 μg/L; Iron <10 μmol/; CRP >5 mg/L
Renal disorder	eGFR <30 mL/min/1.73 m²	eGFR <30 mL/min/1.73 m²
Possible renal disorder	30 ≤ eGFR ≤ 45 mL/min/1.73 m²	30 ≤ eGFR ≤ 45 mL/min/1.73 m²
Vitamin B12 deficiency	Vitamin B12 < 138 pmol/l	Vitamin B12 < 138 pmol/l
Possible vitamin B12 deficiency^#	138 ≤ Vitamin B12 ≤ 225 pmol/l	138 ≤ Vitamin B12 ≤ 225 pmol/l
Folate deficiency	Folate <7.0 nmol/l	Folate <7.0 nmol/l
Possible bone marrow disorder	Reticulocytes <30 × 10⁹/l Thrombocytes <400 × 10⁹/l	Reticulocytes <25 × 10⁹/l Thrombocytes <400 × 10⁹/l
Possible haemolytic disorder	Reticulocytes >90 × 10⁹/l/l LDH >248 U/l	Reticulocytes >85 × 10⁹/l LDH >247 U/l
Chronic disease or infection or inflammation	CRP >5 mg/L or ESR >10 mm/hr (age 18–44)/ESR >15 mm/hr (age ≥45); Iron <14 μmol/l; 2.0 ≤ Transferrin ≤ 4.1 g/L or CRP >5 mg/L or ESR >10 mm/hr (age 18–44)/ESR >15 mm/hr (age ≥45); 14 ≤ Iron ≤35 μmol/l; Transferrin <2.0 g/L or CRP ≤ 5 mg/L or ESR ≤ 10 mm/hr (age 18–44)/ESR ≤ 15 mm/hr (age ≥45); Iron <14 μmol/l; Transferrin <2.0 g/L or CRP >5 mg/L or ESR >10 mm/hr (age 18–44)/ESR >15 mm/hr (age ≥45); Iron <14 μmol/l; Transferrin <2.0 g/L	CRP >5 mg/L or ESR >15 mm/hr (age 18–44)/ESR mm/hr > 25 (age ≥45); Iron <10 μmol/l; 2.0 ≤ Transferrin ≤ 4.1 g/L or CRP >5 mg/L or ESR >15 mm/hr (age 18–44)/ESR >25 mm/hr (age ≥45); 10 ≤ Iron ≤25; Transferrin <2.0 g/L or CRP ≤ 5 mg/L or ESR ≤ 15 mm/hr (age 18–44)/ESR ≤ 25 mm/hr (age ≥45); Iron <10 μmol/l Transferrin <2.0 g/L or CRP >5 mg/L or ESR >15 mm/hr (age 18–44)/ESR >25 mm/hr (age ≥45); Iron <10 μmol/l; Transferrin <2.0 g/L
Undetermined cause	None of the above	None of the above

*Likely iron deficiency indicates that iron deficiency is the most likely cause of anaemia even if the patient has values within the normal reference range. ^#Vitamin B12 gray area indicates that functional vitamin B12 deficiency cannot be excluded, although the vitamin B12 levels are within the lower reference range. In these cases the suggestion was given to confirm the clinical suspicion of a vitamin B12 deficiency by methylmalonic acid testing (29). CRP, C-reactive protein; eGFR, estimated glomerular filtration rate; ESR, erythrocyte sedimentation rate; LDH, lactate dehydrogenase.

Statistical analysis

Differences in patient populations between the groups were compared using the Mann–Whitney U test (for the continuous variable Hb) and the Chi-squared test (for categorical variables). Differences in test frequencies and found causes of anaemia were compared using the Chi-squared test. Data were analysed using SPSS version 28. A p-value <0.05 was considered statistically significant.

Results

Study population

Patient characteristics of the study groups are shown in Table 2. There were no clinically relevant significant differences in patient characteristics between the study groups besides a higher number of second requests in the second time period.

Table 2.

Patient characteristics, depending on study group.

Characteristics		Pre-control n = 899	Control n = 810	Pre-pilot n = 566	Pilot n = 422	Significance
Time period		Nov 2018–March 2020	Nov 2020–March 2022	Nov 2018–March 2020	Nov 2020–March 2022
Male (%)	Total	383 (42.6)	359 (44.3)	260 (45.9)	183 (43.4)	ns
Female (%)	Total	516 (57.4)	451 (55.7)	306 (54.1)	239 (56.6)	ns
Male (%)	18–44 years	38 (9.9)	39 (10.9)	29 (11.2)	17 (9.3)	ns
	45–64 years	109 (28.5)	73 (20.3)	53 (20.4)	39 (21.3)
	65–79 years	127 (33.2)	122 (34.0)	96 (36.9)	71 (38.8)
	>79 years	109 (28.5)	125 (34.8)	82 (31.5)	56 (30.6)
Female (%)	18–44 years	201 (39.0)	171 (37.9)	106 (34.6)	98 (41.0)	ns
	45–64 years	153 (29.7)	104 (23.1)	83 (27.1)	65 (27.2)
	65–79 years	71 (13.8)	79 (17.5)	47 (15.4)	33 (13.8)
	>79 years	91 (17.6)	97 (21.5)	70 (22.9)	43 (18.0)
Hb median (IQR)	Male	8.0 (7.4–8.3)	8.0 (7.5–8.3)	7.9 (7.2–8.2)	8.0 (7.3–8.3)	ns
Hb median (IQR)	Female	7.1 (6.6–7.3)	7.0 (6.3–7.3)	7.1 (6.7–7.3)	7.0 (6.5–7.3)	<0.05
Number of GPs		30	30	22	23	na
Median amount of requests per GP (IQR)		21 (15–36)	25 (11–41)	28 (17–35)	20 (11–25)	ns
Amount of second requests (%)		100 (11.1)	161 (19.9)	89 (15.7)	85 (20.1)	<0.05

IQR = interquartile range 25th percentile to 75th percentile. A Chi-Square test or Mann–Whitney U test was used to determine if differences between the groups were statistically significant. ns, non-significant (p > 0.05). na, not applicable.

Test frequencies

Table 3 shows the frequencies of performed tests in patients with newly diagnosed anaemia. No differences were observed at baseline (i.e. pre-control and pre-pilot group). Patients within the control group were slightly more likely to receive blood tests than patients in the pre-control group. The (CDS-anaemia) pilot group showed the largest increase in test frequencies. The percent increase that was found to be significant (p < 0.05) ranged from 6.3% for leukocytes, 7.4% for thrombocytes, 44.7% for vitamin B12, 48.9% for iron/transferrin, 57.7% for folate, 62.4% for LDH, 63.6% for reticulocytes to 67.2% for ferritin. Only creatinine/eGFR and ESR were not found to be increased.

Table 3.

Test frequencies of determined tests per group.

Parameter	Pre-control n = 899	Control n = 810	Pre-pilot n = 566	Pilot n = 422
Ferritin (%)	304 (33.8)	373 (46.0)*	172 (30.4)	412 (97.6)*^#
Iron/Transferrin (%)	144 (16.0)	98 (12.1)*	93 (16.0)	274 (64.9)*^#
Vitamin B12 (%)	329 (36.6)	336 (41.5)*	210 (37.1)	345 (81.8)*^#
Folate (%)	157 (17.5)	93 (11.5)*	84 (14.8)	306 (72.5)*^#
eGFR (%)	672 (74.7)	616 (76.0)	458 (80.9)^#	360 (85.3)^#
Leukocytes (%)	800 (89.0)	760 (93.8)*	505 (89.2)	403 (95.5)*
Thrombocytes (%)	773 (86.0)	751 (92.7)*	496 (87.6)	401 (95.0)*
Reticulocytes (%)	21 (2.3)	40 (4.9)*	33 (5.8)^#	293 (69.4)*^#
LDH (%)	44 (4.9)	46 (5.7)	25 (4.4)	282 (66.8)*^#
CRP (%)	294 (32.7)	274 (33.8)	222 (39.2)^#	286 (67.8)*^#
ESR (%)	367 (40.8)	377 (46.5)*	287 (50.7)^#	201 (47.6)

A Chi-Square test was used to determine if differences in groups were statistically significant. *p < 0.05 control or pilot group compared to the matching pre-group. ^#p < 0.05 pre-pilot compared to pre-control group, pilot compared to control group.

Causes of anaemia

There were no baseline differences observed in possible causes of anaemia between the pre-control and pre-pilot group, as shown in Table 4. The control group showed slightly higher and statistically significant (p < 0.05) frequencies of iron deficiency (18.8% vs 13.7%), chronic disease, infection or inflammation (1.0% vs 0.3%), renal disorder (4.7% vs 2.6%) and possible vitamin B12 deficiency (12.3% vs 8.3%) as compared to the pre-control group. In the pilot group, all causes of anaemia, except for renal disease, were more likely to be found (Table 4): proven iron deficiency: 31.3% vs 18.8%, likely iron deficiency: 11.4% vs 6.0%, iron deficiency in acute phase: 2.6% vs 0.2%, vitamin B12 deficiency: 4.5% vs 1.9%, possible vitamin B12 deficiency: 16.8% vs 8.7%, folate deficiency: 3.3% vs 0.9% and possible bone marrow disorder: 3.8% vs 0.0%. These increases were statistically different from both the pre-pilot and control group(s). Figure 2 shows the most frequent causes of anaemia for the pilot group. Especially the group with an undetermined cause of anaemia was highly reduced in the intervention group.

Table 4.

Detected possible causes of anaemia per group.

Parameter	Pre-control n = 899 (%)	Control n = 810 (%)	Pre-pilot n = 566 (%)	Pilot n = 422 (%)
Iron deficiency	123 (13.7)	152 (18.8)*	82 (14.5)	132 (31.3)*^#
Likely iron deficiency	33 (3.7)	49 (6.0)*	16 (2.8)	48 (11.4)*^#
Iron deficiency in acute phase	4 (0.4)	2 (0.2)	3 (0.5)	11 (2.6)*^#
Chronic disease, infection, or inflammation	27 (3.0)	8 (1.0)*	11 (1.9)	99 (23.5)*^#
Renal disorder	23 (2.6)	38 (4.7)*	22 (3.9)	10 (2.4)^#
Possible renal disorder	55 (6.1)	56 (6.9)	34 (6.0)	29 (6.9)
Vitamin B12 deficiency	11 (1.2)	17 (2.1)	11 (1.9)	19 (4.5)*^#
Possible vitamin B12 deficiency	75 (8.3)	100 (12.3)*	49 (8.7)	71 (16.8)*^#
Folate deficiency	9 (1.0)	8 (1.0)	5 (0.9)	14 (3.3)*^#
Possible bone marrow disorder	2 (0.2)	0 (0.0)	0 (0.0)	16 (3.8)*^#
Possible haemolytic disorder	0 (0.0)	1 (0.1)	1 (0.2)	3 (0.7)
Undetermined cause	607 (67.5)	448 (55.3)*	366 (64.7)	105 (24.9)^A*^#

A Chi-Square test was used to determine if differences between two groups were statistically different. There were no significant differences observed between the pre-control and pre-pilot group. *p < 0.05 control or pilot group compared to the matching pre-group. ^#p < 0.05 pilot compared to the control group. (A) In this group no cause was determined, however, all described causes were investigated and were not likely to be causing the anaemia.

Figure 2.

Detected possible causes of anaemia in the pre-pilot and pilot group. The numbers and significance are shown in Table 4.

Discussion

This study investigated the diagnostic yield of the cause of anaemia by comparing an automated algorithm performing guideline anaemia work-up versus manual anaemia work-up by GPs.

Although most countries, including the Netherlands, have guidelines for the work-up of anaemic patients from GPs, the actual implementation of guidelines is disputable and in general there seems to be a noticeable difference between actual guideline adherence and perceived guideline adherence.¹³ For several reasons, overuse of medical services is a general phenomenon.¹⁴ Indeed many GPs choose to add additional tests to the initial laboratory request when having Hb analysed (manuscript in preparation). However, the majority of patients do not have anaemia, therefore over-utilization of laboratory testing is common practice. On the other hand, when anaemia is present only a part of the patients is scrutinized into further detail.^7,8 According to the guidelines of the Dutch Society for Clinical Chemistry and Laboratory Medicine (NVKC), medical diagnostic laboratories should offer a standardized protocol for anaemia work-up.¹⁵ Atalmedial therefore implemented clinical decision support (CDS-anaemia) software to comply with the NVKC guidelines, while following the DCGP guideline.^9,15 Since very little is published about the effectiveness of such automated diagnostic algorithms, we studied the effectiveness of the automated anaemia algorithm.

Most importantly, this study clearly showed that within patients of the pilot (CDS-anaemia) group much higher proportions of patients were diagnosed with an underlying cause of anaemia. On the other hand, the proportion of patients with an undetermined cause dropped from 67.5% in the pre-control group, to 24.9% in the pilot group. Importantly, in the pilot group in the group of patients in which no cause of anaemia could be found, nutrient deficiency, chronic disease, infection and renal failure as a cause of the observed anaemia were excluded by the added tests, whereas in the pre-pilot group these possible causes of anaemia were not consequently investigated. Many more patients were diagnosed with proven, likely or probable nutrient deficiency in the CDS-anaemia group as compared to the control group. This is a very relevant finding since in general nutrient deficiency can be reversed by targeted nutrient supplements or dietary advice, or can direct at somatic disease such as blood loss/gastrointestinal cancer, inflammatory bowel disease or pernicious anaemia.¹⁶

The biggest difference in observed cause of anaemia was found for chronic disease or infection or inflammation (23.5% vs 1.9%). Since improvement of Hb is highly dependent on the type of disease, the laboratory report clearly pinpoints the probable cause and therefore supports the GP in the diagnosis. Despite the low incidence, CDS-anaemia did also aid in the detection of possible bone marrow disorders (3.8% vs 0.0%). Although the follow up of these patients was not part of this work, bone marrow disorders, especially myelodysplastic syndromes are more common in the elderly and therefore future research should address this finding into more detail.¹⁷ Chronic kidney disease as a cause of anaemia was not found to be increased when using CDS-anaemia. This can easily be explained by the notion that kidney disease is highly correlated with age and GPs are used to measure creatinine to determine renal clearance in the elderly.¹⁸ Therefore CDS-anaemia does not seem to aid in the detection of more kidney related anaemia.

To be able to increase the number of causes found, CDS-anaemia consequently adds laboratory tests, when Hb is below reference value. The most striking increase for nutrients deficiency was disclosed by ferritin testing. The proportions for testing frequencies went up from 30.3% pre-CDS-anaemia to 97.2% during implementation. This meant that CDS-anaemia was very effective at adding ferritin and pointing to iron deficiency as the most likely cause of anaemia, according to the anaemia guideline of the DCGP. Although the DCGP anaemia guideline does not include ferritin testing for women in their reproductive age, we found that many GPs particularly favoured ferritin testing for this group.⁸ The small number without ferritin was caused by the lack of the necessary testing material. The clinical relevance of increased test frequencies for blood parameters is important to consider. Literature indicates that the ability of GPs to diagnose a specific cause for anaemia increases with more blood diagnostics.^19,20 The use of an automated-anaemia algorithm supports this implication, so more causes of anaemia are found in the CDS-anaemia group. This may eventually lead to a more favourable outcome for the patient, since specific utilization of diagnostics can accelerate targeting the disease and shorten the time in anaemia or distress.

The unexplained anaemia results from our pre-control (67.5%), control (55.3%) and pre-pilot (64.7%) groups are similar to the results of Droogendijk et al. but exceed current literature that unexplained anaemia ranges around 28–30%.^7,21,22 The simplest explanation is lack of follow-up testing in the anaemic individuals.⁸ CDS-anaemia reduced this number of unexplained anaemia to 24.9%, in line with the findings of unexplained anaemia in the elderly.

The anaemia algorithm implemented was based upon the DCGP guideline anaemia.⁹ However, some aspects of the implemented algorithm can be criticized and some improvements could potentially be made. Firstly, do we really need LDH with each vitamin B12 and folate measurement and in which cases do we need reticulocytes? Also the interpretation is quite complex for certain less common causes of anaemia such as haemolytic disease. Secondly, the guideline and algorithm use a one size fits all approach. We, however, know that in different patients groups different conditions cause anaemia.^8,9 Therefore, an algorithm based upon the differential diagnosis per patient group would be more cost-effective. In addition, a step-by-step approach could be an option, although potentially some patients may have multiple causes of anaemia, which might be missed as the analysis stops after one cause of anaemia is found. Both subjects are an important subject of further research and could reduce the amount of unnecessary added tests.²³ Lastly, it is our experience that in a significant number of patients anaemia related tests are requested together with Hb. Since the majority of the Hb tests appear to be non-anemic,²⁴ the algorithm can be helpful by reducing unnecessary laboratory testing and ultimately improve diagnostics and reduce costs at the same time.

Future research should focus on what value CDS-anaemia really adds to anaemia care. Since it is unknown if GPs actually diagnose more causes of anaemia with CDS-anaemia assistance, future research should evaluate that topic. Even more importantly, it is also unknown whether or not patients receive more adequate treatment with CDS-anaemia assistance. This requires more intensive research, since treatment originating from the GP needs to be quantified. And lastly it is not yet investigated whether CDS-anaemia leads to a better clinical outcome of anaemic patients.

Another topic of investigation should include the financial impact blood tests have on diagnosing anaemia. Kip et al. showed by an online survey that extensive testing by GPs leads to increased cause finding of anaemia and consequently increase in costs.¹⁸ Although the increase in costs was reported to be limited, it may be advantageous to investigate whether the (DCGP) anaemia algorithm can be tuned to optimize effective blood testing. Alternative approaches have been suggested by Oosterhuis et al.²³ but state of the art data mining and logistic solutions may help to proceed in this field.

Conclusion

The results of this study show that the cause of anaemia in anaemic patients of GPs can be determined in a larger proportion of patients when GPs are supported by an automated algorithm performing guideline anaemia testing than in anaemic patients of GPs who have to do the work-up manually.

Footnotes

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) received no financial support for the research, authorship, and/or publication of this article.

Ethical approval

N/A.

Guarantor

BdB.

Contributorship

All authors reviewed and edited the manuscript and approved the final version of the manuscript.

ORCID iDs

Anne Margreet de Jong

Firmin Candido

Bauke A de Boer

References

World Health Organization . Worldwide prevalence of anaemia 1993-2005. In: Benoist

McLean

Egli

, et al. (eds) WHO global database on anaemia. World Health Organization; 2008. https://apps.who.int/iris/handle/10665/43894

Stevens

Finucane

De-Regil

, et al. Global, regional, and national trends in haemoglobin concentration and prevalence of total and severe anaemia in children and pregnant and non-pregnant women for 1995-2011: a systematic analysis of population-representative data. Lancet Glob Health 2013; 1: e16–e25.

Safiri

Kolahi

Noori

, et al. Burden of anemia and its underlying causes in 204 countries and territories, 1990-2019: results from the global burden of disease study 2019. J Hematol Oncol 2021; 14(1): 185.

Guralnik

Eisenstaedt

Ferrucci

, et al. Prevalence of anemia in persons 65 years and older in the United States: evidence for a high rate of unexplained anemia. Blood 2004; 104(8): 2263–2268.

Cascio

DeLoughery

. Anemia: evaluation and diagnostic tests. Med Clin North Am 2017; 101(2): 263–284.

Kassebaum

GBD 2013 Anemia Collaborators . The global burden of anemia. Hematol Oncol Clin North Am 2016; 30(2): 247–308.

Droogendijk

Berendes

Beukers

, et al. Adherence to the Dutch general practitioner anemia guideline a retrospective cohort study. Perfusion (Germany) 2013; 26: 184–190.

De Jong

Bozic

Candido

, et al. Onderdiagnostiek bij onderzoek naar anemie in de eerste lijn. Ned Tijdschr Geneeskd 2022; accepted for publication.

Anemie . NHG-Richtlijnen. https://richtlijnen.nhg.org/standaarden/anemie (2014, accessed 25 March 2022).

10.

Kip

MMA

Oonk

MLJ

Levin

, et al. Preventing overuse of laboratory diagnostics: a case study into diagnosing anemia in Dutch general practice. BMC Med Inform Decis Mak 2020; 20(1): 178.

11.

Zhi

Ding

Theisen-Toupal

, et al. The landscape of inappropriate laboratory testing: a 15-year meta-analysis. PLoS ONE 2013; 8(11): e78962.

12.

Elnenaei

Campbell

Thoni

, et al. An effective utilization management strategy by dual approach of influencing physician ordering and gate keeping. Clin Biochem 2016; 49(3): 208–212.

13.

Lomas

Anderson

Domnick-Pierre

, et al. Do practice guidelines guide practice? The effect of a consensus statement on the practice of physicians. N Engl J Med 1989; 321(19): 1306–1311.

14.

Brownlee

Chalkidou

Doust

, et al. Evidence of overuse of medical services around the world. Lancet 2017; 390(10090): 156–168.

15.

Verboeket-van de Venne

WPHG

Oosterhuis

Keuren

JFW

, et al. Richtlijn NVKC Reflexdiagnostiek bij anemie. https://www.nvkc.nl/sites/default/files/6.pdf (2014, accessed 16 September 2022).

16.

Mole

Iron

Davis B.

vitamin B12 and folate. Medicine 2017; 45(4): 198–203.

17.

Cogle

. Incidence and burden of the myelodysplastic syndromes. Curr Hematol Malig Rep 2015; 10: 272–281.

18.

Zhang

Rothenbacher

. Prevalence of chronic kidney disease in population-based studies: systematic review. BMC Public Health 2008; 8: 117.

19.

Schop

Kip

Stouten

, et al. The effectiveness of a routine versus an extensive laboratory analysis in the diagnosis of anaemia in general practice. Ann Clin Biochem 2018; 55(5): 535–542.

20.

Kip

Schop

Stouten

, et al. Assessing the cost effectiveness of a routine versus an extensive laboratory work-up in the diagnosis of anaemia in Dutch general practice. Ann Clin Biochem 2018; 55(6): 630–638.

21.

Michalak

Rupa-Matysek

Hus

, et al. Unexplained anemia in the elderly – a real life analysis of 981 patients. Arch Med Sci 2020; 16(4): 834–841.

22.

Alvarez-Payares

Rivera-Arismendy

Ruiz-Bravo

, et al. Unexplained anemia in the elderly. Cureus 2021; 13(11): e19971.

23.

Oosterhuis

Van der Horst

van Dongen

, et al. Prospective comparison of the flow chart for laboratory investigations for anaemia from the dutch college of general practitioners' guideline 'Anaemia' with a self-developed, substantive and logistical alternative flow chart'. Ned Tijdschr Geneeskd 2007; 151: 2326–2332.

24.

Netherlands Institute for Health Services Research . Nivel primary care database. https://www.nivel.nl/nl/nivel-zorgregistraties-eerste-lijn/jaarcijfers-aandoeningen-huisartsenregistraties

The effectiveness of an automated algorithm as a tool for investigating the cause of anaemia in undiagnosed patients from general practitioners

Abstract

Background

Methods

Results

Conclusions

Keywords

Introduction

Methods

Study design

Pilot and control group

Patient characteristics

Definitions and anaemia algorithm

Statistical analysis

Results

Study population

Test frequencies

Causes of anaemia

Discussion

Conclusion

Footnotes

Declaration of conflicting interests

Funding

Ethical approval

Guarantor

Contributorship

ORCID iDs

References