WhatsApp in Diabetes Care: What 90 Studies Actually Tell Us

Key Statistics at a Glance

Records Retrieved	Studies Included	Total Citations (WoS)	Causal Chains	Publication Span
135	90	739	5	2017–2026

01 · Why This Question Matters

Every second, over 140,000 messages are sent on WhatsApp. Much of this torrent is perfectly mundane — grocery lists, family photos, voice notes about tonight’s dinner. But a growing fraction of it is something altogether different: people with diabetes discussing insulin doses, sharing blood glucose readings with nurses, receiving reminders to take metformin, or finding emotional support in peer groups organized around living with a chronic condition. The platform was never designed for this. Yet here we are.

This analysis maps 90 peer-reviewed studies — published between 2017 and 2026 — to answer a deceptively simple question: exactly how is diabetes being positioned on WhatsApp as a communication platform?

The answer, it turns out, is more fragmented than optimistic headlines would suggest. Yes, WhatsApp is demonstrably useful in certain diabetes care contexts. But the evidence base carries structural gaps, geographic skews, and measurement inconsistencies that should temper enthusiasm before health systems start recommending it at scale.

Methodology note: A Web of Science Core Collection search (Topic: whatsapp AND full-field: Diabet*) returned 135 records. After excluding 45 studies — for non-diabetes health topics, purely instrumental WhatsApp use (survey distribution only), or peripheral diabetes-platform connection — 90 studies were retained for analysis. A six-dimensional measurement taxonomy and five-chain causal analysis were applied to the final corpus.

02 · The Noise in the Signal

Before asking what the literature says, it is worth asking what the literature is — and what it is not. Of the 135 records retrieved, 45 (33%) were excluded from analysis. This exclusion rate is itself informative.

The largest exclusion category — 22 studies — involves entirely different health conditions where WhatsApp appears only as a data collection channel: varicose veins among surgeons, carpal tunnel syndrome in schoolteachers, osteoporosis in postpartum women, HIV support groups, colorectal cancer quality of life. These studies entered the search results because diabetes appeared somewhere in their keyword enrichment fields, not in their actual content. A second category of 9 studies used WhatsApp exclusively to distribute survey links to diabetic populations — the platform was a mailbox, not a subject of inquiry. A third category of 4 studies showed only contextual or tangential connections.

Table 1 — Exclusion classification (n=45)

Exclusion Category	n	% of Excluded	Representative Examples
Non-diabetes health topic	22	48.9%	Cervical cancer, HIV, varicose veins, carpal tunnel, osteoporosis
WhatsApp as survey instrument only	9	20.0%	Anxiety disorder, healthcare professional education, pre-eclampsia
Peripheral / contextual connection	4	8.9%	PCOS, voice quality in T2DM, obesity gamification, COVID infodemic
Total excluded	35	77.8%	—

03 · A Portrait of the Field

The 90 included studies span 2017 to 2026, with a clear acceleration after 2020. The COVID-19 pandemic did not create WhatsApp-based diabetes research — it supercharged it. Publication counts jumped from 3 in 2019 to 12 in 2021, sustained at 12 in 2022, and surged to 21 in 2024. This trajectory reflects a field moving from curiosity to clinical trial, from pilot to policy proposal.

The disciplinary landscape is equally telling. Endocrinology and Metabolism leads (n=22), followed by Health Care Sciences (n=17), Public Health (n=16), and General Internal Medicine (n=10). Medical Informatics appears fifth (n=7) — a sign that the field is becoming increasingly techno-clinical. The 81 original articles and 6 reviews generated 739 total citations (WoS Core), with a mean of 8.2 and a median of 3.0. The top-cited study, Saqib et al. (2020), alone accounts for 125 citations — 17% of the entire corpus’s citation stock.

Table 2 — Publication trend by year (n=90)

Year	Publications	Notes
2017	2	Field inception
2018	4	Early growth
2019	3	Pre-pandemic baseline
2020	8	COVID-19 lockdown — WhatsApp becomes emergency channel
2021	12	Post-COVID consolidation
2022	12	Sustained high output
2023	10	Continued growth
2024	21	Publication peak — largest single-year output
2025	13	High trajectory continues
2026*	5	*Partial year at time of analysis

04 · Six Dimensions of Measurement

To understand how diabetes is positioned on WhatsApp, we need to look beyond publication counts and examine what researchers are actually measuring. The corpus reveals six distinct measurement dimensions — outcome variables, intervention types, study designs, target populations, geographic contexts, and measurement instruments. Their distribution exposes the architecture of the field.

Table 3 — Measurement taxonomy across six dimensions (n=90)

Dimension	Category	n	Frequency	Interpretation
Outcome variable	Knowledge / awareness	63	70.0%	Dominant; cognitive entry level
Outcome variable	Self-care behaviour	37	41.1%	Behavioural mid-level
Outcome variable	Glycaemic control (HbA1c)	32	35.6%	Key clinical endpoint
Outcome variable	Physical activity	21	23.3%	Behavioural activation
Outcome variable	Psychological wellbeing	19	21.1%	Under-represented patient perspective
Intervention type	Cross-sectional survey	47	52.2%	Still in descriptive phase
Intervention type	WhatsApp direct education	42	46.7%	Core intervention modality
Study design	Randomised controlled trial	30	33.3%	Level II evidence emerging
Study design	Systematic review / meta-analysis	5	5.6%	Synthesis capacity underdeveloped
Population	Type 2 diabetes	37	41.1%	Dominant clinical population
Population	Gestational diabetes	5	5.6%	Critical research gap
Geography	Middle East / Gulf	28	31.1%	Highest diabetes burden + WA dominance
Geography	South / SE Asia	22	24.4%	Together: 55.5% of corpus
Measurement instrument	Clinical biomarker	30	33.3%	Gold-standard objective measure
Measurement instrument	Validated psychometric scale	4	4.4%	Psychometric standardisation absent

“The fact that 70% of studies measure knowledge gains — while only 5.6% attempt evidence synthesis — reveals a field that is widening at the base faster than it can consolidate at the apex.”

05 · Five Routes from Message to Outcome

Content mapping of abstracts, keywords and titles revealed five distinct causal chains — each representing a different theory of how WhatsApp creates value in diabetes management. These are not hypothetical constructs: they are inferred directly from how researchers frame their research questions, design their interventions, and define their primary outcomes.

Chain 1: Knowledge → Glycaemic Control  [Dominant chain]

Low literacy → WA education (n=41) → Knowledge gain (n=48) → HbA1c (n=25)

The largest mechanism node (n=41). Only 3 studies cover all four links. Al Omar et al. (2020), Mash et al. (2022) and Alanzi et al. (2018) are anchors. The missing link: no study systematically measures the knowledge-to-behaviour transition as a unified causal process.

Full-link coverage: Full overlap n=3

Chain 2: Activation → Metabolism  [Narrow bridge chain]

Motivation (n=18) → Pedometer+WA (n=6) → Physical activity (n=21) → BMI/BP (n=14)

Wearable + WhatsApp integration appears in only n=6 studies. Alghafri et al. (2018) MOVEdiabetes RCT is the benchmark. The near-absence of integrated wearable-WA designs represents a critical design gap.

Full-link coverage: Full overlap n=1  ⚠

Chain 3: Telehealth → Care Continuity  [Gap chain]

Access barrier (n=15) → Telemedicine (n=12) → Prof. training (n=16) → Service use (n=23)

Distal node (n=23) is relatively strong, yet no study covers all four links. A structural blindspot especially relevant for low-resource settings where WhatsApp is used as de facto telemedicine infrastructure.

Full-link coverage: Full overlap n=0  ✕

Chain 4: Social Support → Quality of Life  [Lost chain]

Motivation (n=18) → Social media support (n=30) → Psychological (n=18) → QoL (n=9)

Quality of life as distal endpoint has the lowest frequency (n=9). Tang et al. (2024) and Göçen & Öztürk (2025) are rare examples measuring psychological wellbeing as primary outcome. The patient perspective is the most undervalued dimension in the entire corpus.

Full-link coverage: Full overlap n=0  ✕

Chain 5: COVID Crisis → Self-Care → Glycaemia  [Dual-output chain]

Diabetes burden (n=36) → COVID adaptation (n=19) → Self-care (n=33) → HbA1c (n=25)

COVID accelerated WhatsApp adoption as a mandatory channel. Saqib et al. (2020) with 125 citations anchors this chain alongside Sisman et al. (2022). Crisis conditions shaped the design of most studies in this chain — caution required in generalising to planned programme delivery.

Full-link coverage: Full overlap n=2

06 · What We Still Do Not Know

Perhaps the most valuable output of a systematic mapping exercise is not a catalogue of what exists, but a precise account of what is missing. Four critical evidence gaps emerge from this analysis.

Table 4 — Priority research gaps

Gap	Description	Priority
GAP 1 — Integrated chain	No study spans the full low-literacy → WA education → knowledge → HbA1c pathway within a single design. The knowledge-to-behaviour mechanism is unmeasured as a unified process.	High
GAP 2 — Wearable integration	The wearable-WA mechanism node contains only n=6 studies, with four-link overlap at n=1. As wearables enter mainstream diabetes management, this absence is a design-level blindspot.	High
GAP 3 — Telehealth continuity	No single study traces the access barrier → telemedicine → professional training → service utilisation pathway (Chain 3, overlap n=0). Critical for low-resource settings.	High
GAP 4 — Patient-centred outcomes	Quality of life and psychological wellbeing appear as distal outcomes in only 10% and 21.1% of studies. Validated psychometric scales used in just 4 studies (4.4%). The field measures HbA1c rather than patient-reported wellbeing.	Urgent

07 · What This Means in Practice

For a clinician, health manager, or policymaker asking whether to integrate WhatsApp into a diabetes programme, this analysis offers a nuanced answer: it depends entirely on which chain you are trying to activate, and whether you are prepared to measure it properly.

The strongest evidence exists for Chain 1 — WhatsApp as a structured educational intervention improving knowledge and, in some cases, glycaemic control. Sartori et al. (2020), Al Omar et al. (2020), and Mash et al. (2022) represent genuine high-quality trials with measurable clinical effects. If your goal is to improve disease knowledge in populations with low health literacy — particularly in Middle Eastern and South Asian settings where WhatsApp is the dominant mobile communication platform — the evidence is sufficiently robust to inform programme design.

Chain 5 offers a cautionary tale: COVID forced health systems to use WhatsApp as telemedicine infrastructure, and the evidence produced during this period (2020–2022) is disproportionately shaped by crisis conditions. Studies from this era should be interpreted with care when applied to planned, non-emergency programme delivery.

Chains 2, 3, and 4 represent areas where programmes may be running ahead of the evidence. Wearable integration, professional training pathways, and psychological support interventions via WhatsApp are being implemented in various health systems without an adequate evidence base. This is not an argument against innovation — it is an argument for embedding rigorous measurement into every pilot.

The geographic concentration of evidence (55.5% from the Middle East and South/Southeast Asia) means that generalisability to European, North American, or Sub-Saharan African contexts requires explicit contextual adaptation. A WhatsApp group that works brilliantly in a Riyadh diabetes clinic may need substantial re-engineering to function in a rural Nigerian primary care setting or a Swedish telehealth system.

Table 5 — Evidence readiness by causal chain

Chain	Mechanism	Full Coverage	RCT	Scale-Up Readiness
Z1 Knowledge	WA direct education (n=41)	n=3	Yes	Moderate–High
Z5 COVID/Self-care	COVID adaptation (n=19)	n=2	Partial	Moderate (context-specific)
Z3 Telehealth	Telemedicine (n=12)	n=0	Limited	Low–Moderate
Z4 Psychosocial	Social media support (n=30)	n=0	Emerging	Low
Z2 Activation	Wearable+WA (n=6)	n=1	1 RCT only	Very Low

08 · The Bigger Picture

WhatsApp has become, somewhat inadvertently, one of the most widely deployed health communication platforms in the world. Its deployment in diabetes management is not the result of a grand digital health strategy — it is the result of a billion people choosing it as their preferred way to communicate, and health workers adapting to meet them where they are.

That pragmatic origin is both its strength and its limitation. Strength, because low-cost, high-penetration tools that require no new infrastructure can reach populations that formal health systems consistently fail. Limitation, because adoption without rigorous design produces a literature that is broader than it is deep — knowledge-level outcomes measured with unstandardised instruments, geographic concentration in high-prevalence regions, and five causal pathways only two of which have sufficient evidence to guide scale-up decisions.

The next frontier for this literature is not more cross-sectional surveys asking whether patients prefer WhatsApp. It is multi-component RCTs that test integrated chains — connecting wearable sensors, AI-personalised messaging, and psychometric outcome measures in populations that current research consistently underserves

If the trajectory of the past four years — 21 publications in 2024 alone — is any guide, this frontier will be reached sooner than the systematic review literature can keep pace with. The responsibility of researchers working in this space is not just to add data points to an already-crowded chart, but to ensure those data points are connected, comparable, and designed to answer the questions that actually matter for the 537 million people living with diabetes worldwide.

Key References

Al Omar, M., Hasan, S., Palaian, S., & Mahameed, S. (2020). The impact of a self-management educational program coordinated through WhatsApp on diabetes control. Pharmacy Practice-Granada, 18(2), 1841. https://doi.org/10.18549/PharmPract.2020.2.1841

Alghafri, T. S., et al. (2018). MOVEdiabetes: A cluster randomized controlled trial to increase physical activity in adults with type 2 diabetes in Oman. BMJ Open Diabetes Research & Care, 6, e000489. https://doi.org/10.1136/bmjdiab-2017-000489

Alanzi, T., et al. (2018). Evaluation of a mobile social networking application for improving diabetes type 2 knowledge. Journal of Comparative Effectiveness Research, 7(3), 243–254. https://doi.org/10.2217/cer-2017-0048

Chen, S., et al. (2026). Instant messaging-based digital health interventions for diabetes management: A domain-structured meta-analysis. Frontiers in Public Health, 13, 1511785. https://doi.org/10.3389/fpubh.2026.1511785

Mash, R., Schouw, D., & Fischer, A. E. (2022). Evaluating the GREAT4Diabetes WhatsApp chatbot during COVID-19. JMIR Diabetes, 7(1), e32323. https://doi.org/10.2196/32323

Saqib, M. A. N., et al. (2020). Effect of COVID-19 lockdown on patients with chronic diseases. Diabetes & Metabolic Syndrome: Clinical Research & Reviews, 14(6), 1621–1623. https://doi.org/10.1016/j.dsx.2020.08.028

Sartori, A. C., et al. (2020). Educational intervention using WhatsApp on medication adherence in hypertension and diabetes. Telemedicine and E-Health, 26(6), 755–760. https://doi.org/10.1089/tmj.2019.0061

Yaagoob, E., et al. (2024). WhatsApp-based intervention for people with type 2 diabetes: A randomized controlled trial. Nursing & Health Sciences, 26, e13077. https://doi.org/10.1111/nhs.13077