The preanalytic phase, i.e. the series of steps before the actual testing of a laboratory investigation, is known to incur significantly more errors (50-70%) than the analytic (5-15%) and the post-analytic phases (10-30%) [1]. The regulated environment of the laboratory and the implementation of rigorous quality controls have more than decimated the analytic errors. However, the pre-analytic and post-analytic phases are much less regulated, especially at the extreme ends of the process loop (the brain-to-brain loop) [1,2]. A particularly problematic area of the preanalytic phase involves those requiring patient preparation [3].

Fasting is the most well-known type of patient preparation for diagnostic medical testing. Though a multitude of laboratory parameters are significantly affected by fasting state, e.g. glucose, triglyceride, insulin, alkaline phosphatase, bilirubin, iron, C-Reactive Protein, alanine aminotransferase, albumin, calcium, sodium, magnesium, potassium, uric acid, some of the routine hematology count parameters, etc. [4-7] it is glucose for which fasting measurement is most common patient preparation asked by a medical laboratory. Under-fasting and over-fasting, both are detrimental to glucose measurements, for diagnostic as well as monitoring purposes [8].

Fasting requirements are not comprehensively defined in most of the guidelines themselves [9] and are not adequately communicated or explained to patients [10,11]. "Fasting" in laboratory science is not merely a term that is intuitive from its face value and is different from other forms of medical fasting eg “Nil Per Oral (NPO)” before surgery or endoscopic procedures etc. Depending on the test, fasting requires specifying the duration and a list of allowed/desired and prohibited types of activities before and during the fast [4].

For example, the fasting duration for fasting blood glucose may vary from 8-12 hours (usually overnight) depending on country and guideline [4,12]. It also requires 24 hours of alcohol restriction. Any caloric intake in the form of solid or liquid is prohibited during the total duration of fasting. Smoking, exercise, chewing gums, and taking nonessential medication in the morning before sampling is discouraged [13]. However, hydration in the form of clear water is allowed or rather encouraged, and thus it is very different from NPO.

Patients were selected using a convenient sampling method. After obtaining informed consent from the patient, the survey was administered non-judgmentally in the native language of the patients (Hindi or Marwari).

A comparable Online SurveyMonkey tool (SurveyMonkey, San Mateo, CA, USA) [15] was used to survey adult internet users from the same locality, who completed school education or more. This questionnaire was in English.

Mentally incapable, deaf, or dumb patients were excluded from the face-to-face survey. Fully trained Health professionals e.g. doctors, technicians, nurses, etc. were also excluded from both forms of survey.

The questionnaire included questions on

• Basic demographic data including age, gender, locality
• Formal education level
• Financial status (family annual income)
• Perception of hours of fasting required
• Perception about whether to restrict
  • Water intake
  • Beverage
  • Light breakfast
  • Religious eatables e.g. Prasad for Hindu worship or Sehri (pre-dawn meal) for Ramadan etc.
  • Smoking
  • Drinking alcohol
  • Medication
  • Exercise during the fasting
  • Change of, in the previous three days, exercise habit, food
  • Instruction/explanations about fasting received
    • Whether instructed at all
    • Who instructed
    • Verbal or in writing
    • What was the number of hours they fasted in practice
  • Did they take any edibles in the morning
  • What food or beverage they ate before starting the fasting

Table-2: Financial stratification of Patients and Controls.

Perception and practice of fasting duration varied from 0 hours to 17 hours in both groups. A staggering 71% (70/98) of the patients and 35% (65/187) of educated respondents perceived that nobody instructed or explained to them anything about the nature of the fasting.

Table-3: Number (%) of patients, controls, and total, that were instructed or aware of fasting.

Among patients, 11% of the uninstructed knew the fasting duration right, whereas 40% of instructed patients knew the fasting duration right. The instruction did make a significant improvement in awareness about the duration (N=98, Pearson Chi Square=5.321, df=1, p-value=0.021).

Other common sources included another patient (13%), the internet (11%), explanation by phlebotomists after improper fasting (9%), counseling after this questionnaire (6%), and nurses (2%). Top sources for the control were doctors (36%), Internet (20%), this questionnaire (15%), nurses (5%), and previous improper fast (4%).

In the patient group, of those instructed 83% were aware and 72% were compliant in practice about avoiding light snacks during fast. Whereas, among uninstructed patients, 48% claimed to be aware and only 15% were compliant about light snacks. Awareness and compliance about avoidance of tea/coffee was 67% and 44% for instructed vs 30% and 0% for the uninstructed respectively.

An alarming 33% of the instructed patients continued to perceive that religious food could be taken in the morning, and 11% even consumed religious food. As much as 55% of the uninstructed patients perceived religious food was allowable and 19% consumed them. Interestingly 100% of both instructed and uninstructed claimed to be aware of alcohol and smoking. Their compliance was close to 95%.

Even among the controls, 30% thought tea/coffee is allowed, 10% thought snacks are allowed and 3% thought religious food was allowed. Only 35% of controls could guess the correct fasting duration for glucose despite their higher formal educational status.

Among patients, the commonest sources (nonexclusive) of information were: doctors (17% verbal, 2% written, 0% print), another patient (13%), internet (11%), phlebotomist (9%), counselling after this questionnaire (6%). Top information sources for controls were: doctors (36%), Internet (20%), this questionnaire (15%), nurses (5%), and previous improperly prepared tests (3%). Among the instructed controls 83% understood and 72% were compliant about light snacks. Among un-instructed controls, 48% claimed to be aware and 15% adhered to the requirement to avoid snacks.

Awareness and compliance about tea/coffee among patients was 67% and 44% for instructed vs 30% and 0% for the uninstructed. As much as 33% of instructed patients thought religious food could be taken in the morning during fasting, and 11% took them.

Only 15 (15%) patients and 65 (35%) controls recalled fasting for the proper duration. Though the control group was also grossly improper, mostly under-fasting, they had statistically better performance than patients in terms of hours of fasting required. When plotted as a continuous variable, perception of hours of fasting duration needed for glucose was lower than the target (8-12h) in both patients and controls (i.e. majority reported under-fasting), but patients (3.4±4.9h) had a significantly lower duration than controls (5.4±3.7h) and the difference was highly significant (Student’s t-test p-value=0.0004)

Table-6: Fasting Duration Categories in Patients versus Controls.

In the combined population of patients and controls, educational categories were cross-tabulated against fasting duration categories.

Table-7: Fasting duration versus education categories.

Table-8: Proper Fasting duration cross-tabulated against Education categories.

Educational categories had a significant association with better overall knowledge which however was not as significant as the impact of instruction. There was a positive Spearman Correlation of +0.185 between education categories and overall knowledge, which was very significant (p-value=0.002)

In contrast to instruction and educational categories, financial categories did not have any significant association with fasting duration (N=266, Pearson Chi-Square=11.990, df=12, p-value=0.446), but there was a mildly significant association with knowing other conditions right (N=266, Pearson Chi-Square=10.796, df=4, p-value=0.029). On a closer look, the association showed a weak negative Spearman correlation (R=

Though the exact degree of reliance on medical decisions on laboratory investigations is difficult to delineate [17] its criticality is unquestionable. Poor investigation results lead to misdiagnosis and contribute to the burden of morbidity and mortality in many ways. Preanalytic variability, though relatively less standardized than the analytic phase, has several components that have been investigated for decades. However, patient preparation, though the most obvious and unavoidable, is still one of the most neglected components of the pre-analytic phase [18].

India is the “diabetes capital of the World” [19] with 45 million diabetics (15% of the world’s burden) living in this country. Rajasthan is the largest state in India that historically had a very low prevalence of diabetes (<5% in 1990). However, Rajasthan has shown an alarming jump in diabetes prevalence, increasing by 28-36% in the period 1990-2016 [20]. Diagnosis and monitoring of diabetes become all the more important in this context, which critically depends upon the quality of glucose testing, which includes the quality and duration of the fasting.

In this study, it was found that a large majority (71%) of our patients and a third of our educated controls did not remember receiving fasting instructions for glucose. Patients are on average less educated in formal school education (Table 1) but more affluent (Table 2) than the control group. So, as expected, patients had less knowledge or baseline understanding about duration and other conditions of fasting compared to controls and the majority came in improper fasting state quantitatively or qualitatively. Controls, though better by statistical significance, also were not properly prepared (Figure 1 and Table 6).

Education and instruction both make a significant impact and are associated with a better understanding of duration and other conditions.

The primary way to prevent the error would be to improve the content, quality, intensity, and frequency of the information delivery to the patients and updating the training of their caregivers of all levels through as many means as possible.

While the physicians and policymakers involved in developing guidelines formulate the need for fasting and are responsible for systematically increasing public and professional awareness, it is the Laboratory professionals who implement the preanalytic phase and directly ensure the quality. A review of the guidelines of fasting in different countries [4] showed that the guidelines vary: e.g. for fasting glucose testing the USA, UK, and Italy ask for 8 hours of fasting, the only USA specifies that there should be no food or drink except water.

On the other hand, Australia and the Czech Republic recommend 8-10 hours of fasting, and Germany requires 12 hours of fasting overnight. Despite a few newer guidelines [13,23], this field is still far from harmonization. Unless there is detailed research about the bottlenecks in the implementation, policymakers would not know the ground reality of how well the preanalytic guidelines are being implemented and where it needs to be improved.

Some of the limitations of the study include cross-sectional design, inherent limitations of the survey as a tool (e.g. recall bias), and an unusual combination of the financially affluent but educationally diverse patient population in the catchment area of the hospital. These findings need to be replicated and corroborated and generalized by larger studies in a country like ours, which is heterogeneous in percolation of resources including information availability, health care facilities, and training of personnel as well as the level of education of the local population. The high apparent impact of instruction in patients was because the baseline level of knowledge was low. Also, because of the relatively affluent financial status of the average patients and the relatively small sample size, the current study did not have sufficient representation of the very poor or most underprivileged segment of the population in the Indian context.

There remains an unmet need to improve awareness on this topic at multiple levels, that needs to be done multi-modally including printed leaflets, posters, and audio-visual aids in addition to verbal reminders.

Laboratories also are supposed to have updated, clear, and understandable written instructions which unfortunately are rarely created or maintained. There should be clear policies for sample acceptance rules for fasting samples. “No-sample is better than a bad-sample” principle should be strictly enforced and re-training of the laboratory personnel should include an emphasis on this.

The current study raises an alert that the present system of information delivery to patients is not able to achieve its intended goal. Revised recommendations should be more harmonized and include precise and practically executable requirements for patient preparation. For fasting blood glucose testing, the time of the day (7-9 AM), the exact range of fasting duration (8-12hours), and water consumption ad libitum, should be specified. At least verbal explanations to avoid religious snacks during fasting should be provided to help clarify the difference between medical fasting and religious fasting. Alcohol avoidance for 24 h, abstinence from cigarette smoking and avoiding exercise in the morning, and strict restriction of calorific or stimulating beverages including tea, coffee, etc. should be promulgated.

Dr. Asmita Hazra: Study conceptualization and design, data collection

Dr. Saptarshi Mandal: Analysis and interpretation

Dr. Seema L. Jawalekar: Review and authorization

Dr. Jairam Rawtani: Review and authorization

Ms. Minal Marlecha: Data collection