Compliance and barriers to self-monitoring of blood glucose in patients with gestational diabetes mellitus: A systematic review

Introduction

Gestational diabetes mellitus (GDM) is a common medical complication of pregnancy that can adversely affect the health of both the mother (hypertension, increased chances of cesarean section, hydramnios, preeclampsia, etc.) and the fetus (restriction of fetal growth, macrosomia, hyperbilirubinemia, hypoglycemia, and shoulder dystocia).^[1-3] GDM is characterized by hyperglycemia that is diagnosed for the 1^st time in pregnancy (in the absence of any pre-existing Type-1 or type-2 diabetes).^[3,4] The hyperglycemic situation in GDM is temporary and generally normalizes once the pregnancy is over.^[5,6]

Self-monitoring of blood glucose (SMBG) is often recommended as a part of GDM management.^[7-9] SMBG is performed by putting a drop of blood (obtained by puncturing the skin of the tip of the finger using a needle) on a strip which is then read using a glucose reading meter.^[10]

Despite the emphasis on SMBG-testing as a treatment component of GDM,^[7-9] much is not known about its barriers and compliance among the GDM patients. Most studies related to the barriers and compliance of SMBG-testing are primarily based on Type 1 and Type 2 diabetes patients.^[10-14]

Therefore, to appreciate the existing evidence (and its rigor) of the barriers of SMBG and the factors that determine the compliance to SMBG in GDM patients, this systemic review was done.

In this study, the terms adherence and compliance are used interchangeably due to their similar type of meaning.^[15-17]

Methods

This systemic review does not have a pre-published protocol. This study made every possible attempt to adhere strictly to the PRISMA reporting system to report its findings.^[18]

Eligibility criteria: Inclusion criteria: (1) Studies that assessed the barriers to SMBG-testing and/or the factors attributing to the compliance of SMBG-testing in GDM patients (2) studies that were conducted and published between March 2010 and June 2018, (3) studies of following types were included in the study – qualitative studies, observational studies, experimental studies, and systemic review with or without meta-analysis (that include studies conducted and published in March 2010 or later), and (4) articles that are available in English language only.

Results

In total, 42 papers were identified using the above-described literature search. There were 17 duplicates which were manually screened and excluded. After excluding the duplicates, 25 studies were screened by reading the abstracts (against the eligibility criteria of this study) and 12 studies were selected for full paper review. Finally, six papers were rejected, and the remaining six papers^[24-29] were accepted for this review [ Diagram 1]. Inclusion and exclusion of the studies were strictly based on the eligibility criteria mentioned above.

Close

Diagram 1:

PRISMA 2009 flow diagram^[18]

Export to PPT

Data extracted from the selected studies in this review have been depicted in a summary table [Table 1]. Out of the six papers, four were quasi-experimental studies^[26-29] and two were qualitative studies.^[24,25]

Table 1: Summary table

The risk of bias of the selected studies was assessed using the respective JBI tools available for these types of studies.^[30,31] The critical appraisal depicted that the qualitative studies [Table 2] were of good quality.^[24,25] On the other hand, the critical appraisal of the quasi-experimental studies^[26-29] revealed that only one study^[29] had a relatively good study design, methodology, and analysis [Table 3].^[26-28] It is the only quasi-experimental study^[29] that compared its finding with a control group. None of the quasi-experimental studies^[26-29] did any pre- and post-intervention comparison of their findings.

Table 2: Qualitative studies[31]

Table 3: Quasi-experimental studies[30]

The qualitative studies reviewed in this study identified some of the potential barriers of SMBG in GDM patients. The study conducted in China (n = 17) identified the poor understanding of SMBG-testing among GDM patients and lack of SMBG-recommendation by the health-care providers as the barriers to perform SMBG,^[25] whereas the study conducted in Thailand (n = 30) found fear and worry about performing an SMBG-test as a barrier.^[24]

The results of the quasi-experimental studies reveal the following finding. Cosson et al.^[26] (n = 94) reported that poor compliance to SMBG (n = 35, 38%) was independently associated with family history of Type-2 diabetes in first-degree relatives (OR = 0.38, 95% confidence interval [CI] 0.15–0.98), P = 0.044).^[26] The study further identified that French ethnicity and having medical insurance (for the impoverished patients) were also associated with non-compliance to SMBG in GDM patients (however, the OR and CI were not reported for these findings).^[26] The study by Peleg et al.^[29] unveiled that compliance to SMBG-testing improved by the use of smartphone-based reminders in the participants of the intervention group (t-statistics 2.166, P = 0.0312).^[29]

The remaining quasi-experimental studies measured the compliance to SMBG testing among GDM patients in percentages.^[27,28] The study by Mackillop et al. (n=104) found that the use of a smartphone to record blood glucose readings resulted in good SMBG-testing compliance.^[27] Almost 85% of the participants (46 out of 54) of this study did the test as per the research team’s pre-determined optimum SMBG frequency,^[27] whereas the Sousa et al. study^[28] (n = 135) reported 97.5% of the study participants (119 of 122) to be compliant to the SMBG recommendations. The study also found that about 17% of the GDM patients (21/122) perceived the lancet of the needle (to obtain blood for SMBG) and the time of conducting the SMBG-test as the barriers to SMBG.^[28]

A forest plot or meta-analysis was not considered as a part of this systemic review, as the studies included for this review were not comparable (being a combination of qualitative and quantitative) and the quantitative ones^[26-29] were also quite a few in number.^[18]

Overall, the results obtained from the reviewed studies lacked rigor and external validity. The strength and weaknesses of these studies are enumerated in the successive paragraphs.

Study design and study methodology-wise out of the three quasi-experimental studies that reported good compliance to SMBG-testing,^[27-29] the evidence is stronger only for the Peleg et al. study^[29] since it used a control group to compare its finding (along with an appropriate statistical test). However, the methodology of calculating the compliance to the SMBG-testing was different among these three studies and hence not comparable.^[27-29]

Both of the qualitative studies^[24,25] and two of the quasi-experimental studies^[26,28] addressed the barriers to SMBG testing (mentioned above). However, the evidence is perhaps weakened due to the following reasons. Several components that are not clear in the qualitative studies are – the influence of researchers on the study population,^[24,25] how the cultural or theoretical beliefs of the researchers were related to their study,^[24,25] the language in which the interviews were conducted,^[24] and the quality and validity assessment of the English language translations of the verbatim transcripts.^[25] Whereas the hazy areas of the quasi-experimental studies^[26,28] that dealt with SMBG-barriers were – absence of a control group or a pre-intervention finding to compare with the post-intervention findings,^[26,28] the measurement methods of barriers to SMBG-testing in GDM patients were not consistent among the two studies,^[26,28] and lack of validity-related information of the questionnaire used in one of these studies (to understand the barriers of SMBG in GDM patients).^[28]

None of the studies except Cosson et al.^[26] utilized any diagnostic criteria to define GDM in their study population.

Overall, the studies^[24-29] measured or analyzed what they intended to measure or analyze.

The additional search in the PROSPERO website (https://www. crd.york.ac.uk/prospero/)^[19] (as of June 30, 2018) identified a Cochrane protocol that is in “review ongoing” status.^[32] This study is currently assessing adherence to SMBG-testing in GDM patients as a secondary outcome in randomized control trials, quasi-randomized control trails, and cluster-randomized control trials.^[32] However, if the barriers to SMBG-testing among GDM patients are going to be reviewed or not in this Cochrane review is not clear.^[32]

None of the hand searched literature matched the eligibility criteria of this review, hence not included.

Discussion

The literature review conducted in this study identified a severe shortage of studies that have tried to understand the barriers and compliance to SMBG in GDM patients. There are several shortfalls of the existing literature which are discoursed here. Overall, the study population-related description such as the previous history of GDM, the gestational age at which the study participants were recruited, the obstetric history (the gravida and parity), and age range of participants was not available throughout these studies consistently. Study population characteristics such as previous history of GDM, gravida, and parity-related information were available in only one of the studies done by Youngwanichsetha and Phumdoung.^[24] Only two studies (the qualitative ones) mentioned the gestational age of their study population.^[24,25] The age range of the study participants was available for only one study,^[25] and only three out of the six studies provided a mean age.^[24,26,29] Exclusion criteria were not clearly stated in most studies.^[24-27,29] This lack of basic study population defining features made comparison across these studies difficult.

Study population-related confusion was also witnessed in one study, between its two phases.^[27] In the first phase, a pilot testing of the role of mobile technology in SMBG compliance was done.^[27] Based on the feedback from the participants of the pilot test necessary changes was made to the mobile technology and was re-tested in GDM patients for SMBG compliance in the second phase.^[27] The reporting of the result was unclear because it reported compliance (85%, i.e. 46 out of 54) based on all participants of both phases of the study although the mobile technology in the second phase (upgraded) was not same as the first phase.^[27] The comparison between the two phases was complicated further as the info about what medications the participants were on while in the study was only available for those enrolled in the second phase (who were not treated with insulin or metformin).^[27] Nevertheless, a separate group wise frequencies for the compliance was available in the study.^[27]

Regarding the quasi-experimental studies,^[26,28,29] sampling-related information such as sampling method and sample size calculation was grossly unavailable. The study by Mackillop et al.^[27] was an exception in this regard. It recruited all women who attended the diabetes-maternity clinic (in the second phase).^[27] However, the calculation of power or factors that might have strengthened or compromised the statistical power was not available for any of these non-randomized experimental studies.^[24,26-29] Lack of power related information made it difficult to comment on the strength of the evidence produced from the studies.

Then, regarding the confounders, none of the four quasi-experimental studies,^[26-29] except one,^[26] considered for adjusting the confounders. It was also not clear if any other methods such as randomization or matching were used by the quasi-experimental studies^[26-29] to decrease any bias due to confounding. The effects of an intervention are generally at risk of confounding and efforts should be made to identify and control them to avoid undue bias.^[33] Confounding can be controlled by stratification, adjusting, matching, randomization, etc.^[33]

Next, quasi-experimental studies are considered to be of better rigor and quality when they include a pre- and post-intervention comparison and/or a comparison with a control group, rather than without such comparisons.^[34] None of the quasi-experimental studies^[26-29] included in this review mentioned a pre- and post-comparison of their findings. Out of the four quasi-experimental studies,^[26-29] only one study^[29] compared its findings with a control group. Therefore, in regard to comparability of the results, the studies are relatively weak.

Comparability of the four quasi-experimental studies^[26-29] weakened further since the method of assessing compliance among these studies^[26-29] was not consistent. Cosson et al.^[26] considered participants to be compliant with SMBG when at least 80% of the total (pre- and post-prandial) SMBG tests were done,^[26] whereas, Peleg et al.^[29] measured the compliance by dividing the recommended number of SMBG over a time period by the number of SMBG tests done during that period.^[29] On the other hand, the study by Mackillop et al.^[27] ensured compliance when 18 SMBG tests were done per week. In Sousa et al. study,^[28] the compliance determination method was not clear.^[28]

The causes of non-compliance to SMBG among the GDM patients were studied in the Cosson et al. study.^[26] However, only for one cause of non-compliance (the family history of Type 2 diabetes in first-degree relatives) (OR = 0.38, 95% CI 0.15–0.98, P = 0.044) an OR and CI were available, for assessing the strength of association.^[26] Other causes of non-compliance identified by this study (not-belonging to French ethnicity (n = 28, P = 0.048) and having medical insurance dedicated for the impoverished (n = 9, 26%, P = 0.033)^[26] were difficult to interpret because the results were reported in statistical significance (p-value) only (for the respective variables) with no clear mention about the strength of association (like the OR) and the precision of such association (e.g., the 95% CI).^[26,35]

None of the quasi-experimental studies^[26-29] have a clear mention of the raters and inter- and intra-rater reliability about the cause and effect relationship derived statistically.^[30]

Although the qualitative studies^[24,25] were of good quality (fulfilling most of the critical appraisal criteria [Table 2]), certain issues question the overall rigor of these studies. Only one of the two qualitative studies clearly mentioned its sampling method (purposive).^[25] The interview transcripts in Ge et al.^[25] study were translated from Mandarin Chinese language to the English language; however, information about the translation quality and its validity is not clearly depicted.^[25] In the other qualitative study, interviewers were in contact with the participants for a substantial duration – it is not clear from the study if such engagement anyhow affected (biased) the participant responses and if anything was done to address it.^[24]

Information regarding the GDM diagnostic criteria used to identify GDM patients was available only for one study.^[26] It used the IADPSG criteria^[21] to diagnose GDM.^[26]

Conflict of interest and funding related declaration were available for all studies except the study by Sousa et al.^[28]

None of the authors of the studies^[24-29] incorporated in this review were contacted.

Conclusion

To understand the barriers to SMBG and the factors that determine the compliance of SMBG among GDM patients, this systemic review was conducted. This review included relevant articles published between March 2010 and June 2018. There were few studies (six only) that met the eligibility criteria of this review. The studies reviewed in this paper reported the following barriers of SMBG in GDM patients – lack of SMBG-testing recommendation by health-care providers, poor concept, fear, and anxiety about SMBG-testing, and family history of Type 2 diabetes.^[24-26,28] The compliance to SMBG-testing in GDM patients improved with the use of education (SMBG related) and technology (smartphone and mobile-based).^[27-29] However, these results are not comparable across the studies. Moreover, the results are not generalizable due to the respective weaknesses of the studies discussed in this paper. Therefore, based on this review, the current evidence regarding SMBG related barriers and compliance in GDM patients are labeled as weak and inadequate (for an external validity). Henceforth, further research is required on this topic.