Psychometric evaluation of smartphone addiction scale – short version (SAS-SV) among young adults of India

Abstract

Background:

Although smartphones have considerable utility, they also have addiction potential. The early detection of problematic smartphone use (PSU) can have significant implications for managing its psychosocial consequences. Smartphone Addiction Scale – Short Version (SAS-SV), initially developed for South Korean adolescents, has emerged as a reliable measure for adults across countries. However, SAS-SV continues to be used unvalidated in India.

Aim:

To evaluate the psychometric properties of SAS-SV for the Indian adult population.

Materials and Methods:

Content validation of SAS-SV was done by 10 experts, followed by data collection for validation using a cross-sectional design from 434 participants (M_age = 25.4; SD_age = 2.6; 58.8% females). The datasheet consisted of a sociodemographic questionnaire and SAS-SV. Statistical analyses comprised confirmatory factor analysis (CFA), exploratory factor analysis (EFA), reliability analyses, percentiles, and evaluation of sociodemographic variables.

Results:

SAS-SV’s content validity index was 0.93, and item wordings were adjusted after experts’ feedback. CFA did not show good fit indices; hence, EFA was used, which explained 44% of the variance from a unifactorial model. Cronbach’s alpha was 0.85, McDonald’s Omega was 0.86, and test-retest reliability was 0.81. There were no significant PSU differences in gender, marital status, and occupational status. Higher PSU was associated with lower age, lower education, nuclear family, and more hours of smartphone usage.

Conclusion:

The current study established the psychometric properties of SAS-SV for the Indian adult population. SAS-SV can be used for assessment and treatment monitoring of PSU.

Smartphones are a permanent fixture of modern life. They have an impact on how we interact, travel, purchase, handle money, manage time, socialize, and study. Smartphone use has made it simpler and quicker to acquire knowledge. They provide us access to the Internet so that we can research, email, click photos, record videos, scan QR codes, conduct transactions, or keep track of deadlines.[1] A smartphone is a portable computer that combines computational power and a mobile phone into one device.[2] Popular smartphone manufacturers include Samsung, Apple, and OnePlus using the Android or iOS operating systems. Over 5 billion individuals own mobile devices, with smartphones accounting for more than half of these connections.[3]

Due to the pervasiveness of smartphones and the user’s close relationship with them, there are worries about their potential for addiction. According to Goodman,[4] addiction is a condition of problematic behavior marked by loss of control and use despite harm. Griffiths[5] described technology addiction as a non-chemical, human-machine interaction-based behavioral addiction.[5] Kardefelt-Winther et al.[6] proposed two factors for behavioral addiction: significant functional damage or suffering, and persistent use. The American Psychiatric Association[7] identified internet gaming disorder as a behavioral addiction but not smartphone addiction. However, there has been a noticeable growth in the study on problematic smartphone use (PSU) in recent years.

Screening studies estimate that smartphone addiction ranges from just above 0% to 35%, with the most frequent range between 10% and 20%.[8] Prevalence of smartphone addiction was found to be 36.5% among medical students in Pakistan,[9] 31.7% in Tunis,[10] 16.9% in Switzerland,[11] 16% in Korea,[12] and 9.3% in Iran.[13] In India, Dixit et al. reported a 37% prevalence in medical students.[14] Subsequently, Kumar et al. reported smartphone addiction in 44.7% in another sample of medical students.[15] Davey and Davey’s[2] systematic review and meta-analysis revealed an addiction range of 39%–44% among Indian adolescents. A major limitation of these studies was the use of unstandardized questionnaires for diagnostic purposes. In addition, as of now, neither the DSM-5 nor the ICD-11 recognizes smartphone addiction. Owing to the current lack of diagnostic criteria but the need for continued exploration, the authors have decided to refer to smartphone addiction as PSU.

Despite the absence of PSU from diagnostic manuals, it warrants exploration because it has consistently been linked with decreased social skills, emotional intelligence, and empathy; as well as higher social anxiety and interpersonal conflict.[16] It has also been positively correlated with poor psychological health, poor self-esteem, sadness, insomnia, and anxiety, particularly in adolescents and young adults.[17,18] PSU has also been shown to predict impulsivity, impaired attention, and poorer academic progress.[17,19]

Various screening and diagnostic tools have been made over the years to assess PSU. One of the first scales was created by Toda et al.[20] and is called the Cellular Phone Dependence Questionnaire. Subsequently, other measures such as the Mobile Phone Problem Use Scale,[21] Problematic Mobile Phone Use Questionnaire,[22] Smartphone Addiction Proneness Sale for Korean teenagers,[23] Mobile-Use Screening Test[24] were developed. These tools were constructed across different age groups and assessed various PSU domains.

The Smartphone Addiction Scale was made for Korean adults and later shortened to the Smartphone Addiction Scale – Short Version for Korean adolescents (SAS-SV).[25] SAS-SV is a 10-item unidimensional scale standardized on 540 individuals with a mean age of 14.5 years. The Cronbach’s alpha was found to be 0.91. Based on the ROC analysis, the cutoff for boys was 31, and for girls was 33. SAS-SV has been validated for Moroccan adults, French- and Spanish-speaking Belgian adults, and Italian teens and young adults.[26] SAS and SAS-SV lack Indian psychometric analyses but continue to be used for detection and prevalence studies due to their brevity and popularity.[15,27,28,29] Therefore, the current study aimed to validate SAS-SV in the Indian context and explore its demographic correlates.

MATERIAL AND METHODS

Participants

This study included two types of participants: subject experts who participated in the content validation procedure and participants who formed the sample for the validation of SAS-SV.

Expert Participants: The scale was adapted and modified using the International Testing Committee checklist and criteria, modified by Hernández et al.[30] The inclusion criteria was a degree in M. Phil. Clinical Psychology, MD Psychiatry, or M. Phil. Psychiatric Social Work and a minimum of 5 years of clinical experience after attaining the qualification. Recruitment of 10 experts occurred from the various mental health departments of the authors’ host institute, located in Bengaluru, India. The subject experts evaluated the content validity and the linguistic/cultural validity of SAS-SV per item. Responses were collected on a 4-point scale, where 1 = the item is not relevant and 4 = the item is highly relevant, and interpreted with a content validity index. Next, the experts provided suitable wording suggestions.

Validation Participants: Participants were recruited via non-probability sampling with inclusion criteria of English fluency, Indian resident citizens, and owning a smartphone. Participants with a psychiatric diagnosis in the past year were excluded. The information about the study was disseminated on social media platforms, and those who accepted the informed consent were recruited into the study. A total of 524 individuals participated in the study, of which 89 were excluded because they had a psychiatric disorder (self-reported item in the study questionnaire), and one was excluded due to acquiescent responding. Therefore, the final analysis was done on 434 participants. The participants’ ages ranged from 18 to 30 years, with a mean age of 25.4 years (SD = 2.6). The sample consisted of 255 female-identified participants (58.8%) with a mean age of 25.3 years (SD = 2.5) and 179 male-identified participants (41.2%) with a mean age of 25.7 years (SD = 2.8).

Tools

Demographic questionnaire

A self-reported demographic questionnaire included age, gender, education, marital status, family setting, work status, religion, state, psychiatric diagnosis, and the average smartphone use per day.

Smartphone Addiction Scale – Short Version (SAS-SV)[25]

SAS-SV is a 10-item self-report measure for assessing PSU. It uses a 6-point Likert scale response format ranging from “1” (strongly disagree) to “6” (strongly agree). Kwon and colleagues[25] had 540 participants for the initial development and validation study with a mean age of 14.5 years. It correlated strongly with SAS (r = 0.96) and demonstrated high internal consistency (Cronbach’s alpha = 0.91). SAS-SV has a unidimensional factor structure.[25] In the current study, SAS-SV was used post the completion of the subject expert validation and had Cronbach’s alpha of 0.85. Sample item after expert revision includes, “I keep thinking of my smartphone even when I am not using it.”

Procedure

The author’s (Kim Kwon) permission to use SAS-SV for Indian validation was obtained, and the study was approved by the institutional ethics committee (NIMH/DO/BEH.Sc.Div./2021-22/9^th Nov 2021). The study was conducted in two phases: the expert validation phase and the participant validation phase. The 10 identified experts were approached via email consisting of information about the tool and an informed consent form. After consent, a Google Form was emailed consisting of relevance rating and options to modify the item wording as per Indian linguistic relevance. Upon completion of phase 1, phase 2 was started using validation participants. The first section of their Google Forms consisted of an informed consent form. The next section consisted of the sociodemographic details. The third section consisted of the expert-reviewed questionnaire (SAS-SV). The survey form was circulated on social media. A re-administration for test-retest reliability was done on 30 participants at random around the 14^th day from the completion of the first administration.

Data analysis

Data analyses were conducted using SPSS version 23 (IBM, Armonk, NY, USA) and Ωnyx (Oertzen, Brandmaier, and Tsang). CFA was evaluated using the Comparative Fit Index (CFI) of ≥0.95, the Tukey-Lewis Index (TLI) of ≥0.95, and the Root Mean Square Error of Approximation (RMSEA) of ≤0.08 as adjustment indices. As the CFA did not indicate good fit indices, additional components were investigated using an EFA. Maximum likelihood estimation with oblique rotation was used for the extraction of factors. Only items with a factor weight above 0.40 were considered. The internal consistency of the scale was determined using Cronbach’s alpha statistic and McDonald’s Omega. The test-retest reliability was determined using the intraclass correlation coefficient (ICC). Percentile scores were distributed into the 5^th, 10^th, 25^th, 50^th, 75^th, 90^th, and 95^th. The 75^th percentile and above was considered a score indicating a significantly high PSU. Independent samples t-tests, Pearson’s correlation, and one-way analysis of variance (ANOVA) were used to analyze the sociodemographic differences. Cohen’s d, Pearson’s r, and Eta-squared (η²) were used to determine the effect sizes.

RESULTS

The scale-level content validity index (CVI-S) was found to be 0.93 indicating an acceptable level of CVI. Table 1 shows the CVI-I for each item and the CVI-S for the complete scale.

Table 1.

Content validity index

Subject Experts	1	2	3	4	5	6	7	8	9	10	Agreement among Subject Experts	Item-CVI
Item 1	1	1	1	1	1	1	1	1	1	1	10	1
Item 2	1	0	1	1	1	1	1	1	1	1	9	0.9
Item 3	1	1	1	1	1	1	1	1	1	1	10	0.9
Item 4	1	0	1	1	1	1	0	1	1	1	8	0.8
Item 5	1	0	1	1	1	1	1	1	1	1	9	0.9
Item 6	1	0	1	1	1	1	1	1	1	1	9	0.9
Item 7	1	1	1	1	1	1	1	1	1	1	10	1
Item 8	1	0	1	1	1	1	1	1	1	1	9	0.9
Item 9	1	1	1	1	1	1	1	1	1	1	10	1
Item 10	1	1	1	1	1	1	1	1	1	1	10	1
Scale- CVI=0.93

Item CVI=Item Content Validity Index; S-CVI=Scale-Content Validity Index

The expert committee recommended modifications in the sentence structure of items to improve readability and comprehension for the Indian adult population. SAS-SV with the original items and revised items are outlined in Table 2.

Table 2.

Expert review and item suggestions

Item No.	Original items	Revised items
1.	Missing planned work due to smartphone use.	I miss planned work due to smartphone use.
2.	Having a hard time concentrating in class, while doing assignments, or while working due to smartphone use.	I have difficulties concentrating in class, while doing assignments, or while working due to smartphone use.
3.	Feeling pain in the wrists or at the back of the neck while using a smartphone.	I experience discomfort in my wrists, thumbs, eyes, or at the back of the neck due to smartphone use.
4.	Won’t be able to stand not having a smartphone.	I will not be able to tolerate not having a smartphone.
5.	Feeling impatient and fretful when I am not holding my smartphone.	I feel impatient or irritable when I am not holding my smartphone.
6.	Having my smartphone in my mind even when I am not using it.	I keep thinking of my smartphone even when I am not using it.
7.	I will never give up using my smartphone even when my daily life is already greatly affected by it.	I may not be able to give up my smartphone even though my daily life is affected by it.
8.	Constantly checking my smartphone so as not to miss conversations between other people on Twitter or Facebook.	I constantly check my smartphone so as not to miss any conversations on social media (WhatsApp, Instagram, Facebook, Reddit, Twitter, etc.)
9.	Using my smartphone longer than I had intended.	I use my smartphone longer than I intended or planned.
10.	The people around me tell me that I use my smartphone too much.	My family or friends have told me that I use my smartphone too much.

The current study analyzed data from 434 participants aged 18–30 years (58.8% females) with a mean age of 25.4 years (SD = 2.6). Detailed participant characteristics are in Table 3.

Table 3.

Sociodemographic characteristics of the sample

Sociodemographics	n	%/M (SD)
Gender
Male	179	41.20%
Female	255	58.80%
Age
Age (in years)	Range: 18-30	25.4 (2.6)
Males		25.7 (2.8)
Females		25.3 (2.5)
Age groups
18–23 years	82	18.90%
24–30 years	352	81.10%
Marital status
Single	378	87.10%
Married	56	12.90%
Education
Education	434	100%
12th standard	28	6.50%
Undergraduate	125	28.80%
Postgraduate	281	64.70%
Employment
Students	183	42.20%
Working	251	57.80%
Religious affiliation
Hinduism	291	67.10%
Christianity	78	18%
Islam	25	5.80%
Other	40	9.1%
State of residence (past year)
Karnataka	150	34.60%
Maharashtra	63	14.50%
Madhya Pradesh	45	10.40%
Other	176	40.5%
Smartphone usage time (average per day)
<3 hours	103	23.70%
3–4 hours	132	30.40%
4–5 hours	102	23.50%
>5 hours	97	22.40%
Smartphone Addiction Scores
SAS-SV	Likert-type score: 1–6	29.4 (10)
	Minimum: 10
	Maximum: 60
Male		29.5 (9.9)
Female		29.2 (10.2)

n=Number; %=Percent; M=Mean; SD=Standard Deviation; SAS-SV=Smartphone Addiction Scale – Short Version

The item-wise means and standard deviations for the revised SAS-SV are outlined in Table 4.

Table 4.

Item score analysis of modified SAS-SV

Items	Mean	SD
Item 1	2.85	1.51
Item 2	2.91	1.55
Item 3	2.97	1.66
Item 4	3.44	1.62
Item 5	2.30	1.35
Item 6	2.15	1.32
Item 7	2.74	1.54
Item 8	3.32	1.61
Item 9	3.88	1.50
Item 10	2.80	1.61

SD=Standard Deviation

The fit parameters of the CFA unifactorial model were inadequate (CFI = 0.82; TLI = 0.82; RMSEA = 0.12). Therefore, an approach like the one taken by Andrade and colleagues[31] was employed, where an EFA was conducted to explore if SAS-SV had more than one factor. All the items attained a factor loading above 0.40 (range = 0.49–0.70). The EFA analysis revealed two factors; however, there was a large drop between the variance explained by the first factor, that is, 44%, and the second factor, that is, 13%. In addition, the first factor explained all 10 items adequately. Hence, the unifactorial structure of SAS-SV was retained. The corrected item-total correlation ranged from 0.46 to 0.63. No item was dropped because they maintained an adequate factor loading (>0.40) and retained the overall internal consistency. The factor analysis is depicted in Table 5.

Table 5.

Analysis of factor structure and item-total correlations

Item	Factor loading	Corrected item total r	α if item deleted
1	0.59	0.52	0.84
2	0.60	0.55	0.84
3	0.49	0.47	0.85
4	0.53	0.46	0.85
5	0.70	0.61	0.84
6	0.70	0.62	0.84
7	0.68	0.63	0.83
8	0.63	0.58	0.84
9	0.64	0.59	0.84
10	0.63	0.59	0.84
Eigenvalue	4.40
Variance (%)	44%

r=coefficient of correlation. The extract method was performed using the maximum likelihood method. A cut-off of 0.40 was used for inclusion. Bartlett’s test of sphericity: P<0.001; KMO overall: 0.87. The overall Cronbach’s α of SAS-SV was 0.85

The Cronbach’s alpha coefficient was found to be 0.85, and the McDonald’s Omega was found to be 0.86. The test-retest reliability was found to be 0.81 (P < 0.001). Overall, the scale had good internal consistency and test-retest reliability.

The percentile scores of SAS-SV were calculated gender-wise (male and female), and these are outlined in Table 6. As per Simpson,[32] age segregation was done into two categories: 18–23 years and 24–30 years. The combined percentile (with both genders) was also calculated. The 75^th percentile was considered to indicate significantly high PSU, with higher percentiles indicating higher levels of PSU.

Table 6.

SAS-SV percentiles across gender and age ranges

Percentile	Age range (in years)						Combined percentiles
	Male		Female		Gender combined		(Gender and age)
	18–23	24–30	18–23	24–30	18–23	24–30	Overall
5	14.00	14.00	15.70	12.50	15.15	13.00	14
10	18.10	16.00	18.00	15.00	18.30	16.00	16
25	26.00	21.00	24.00	21.50	24.75	21.00	22
50	35.00	29.00	30.50	29.00	32.00	29.00	29
75	39.00	35.00	41.50	34.00	40.25	34.00	36
90	42.90	41.00	48.50	41.00	46.00	41.00	42
95	46.00	48.80	52.65	45.00	51.25	47.35	47.25

Use the following percentile table to interpret the scores of SAS-SV-Indian Version. Higher scores and higher percentiles indicate higher problematic smartphone use

There was no significant effect of gender (t (432) = 0.35, P = 0.73), marital status (t (432) = 1.6, P = 0.11), and working status (t (432) = 0.78, P = 0.44) on PSU. Age had an inverse relationship with PSU (r = −0.14, P < 0.01). Scores on SAS-SV of those educated up to the 12^th standard (M = 32.7, SD = 10.1) and graduation (M = 32.3, SD = 10.4) were significantly higher than compared to those with post-graduation and above (M = 27.7, SD = 9.6) with a small effect size, ƞ² = 0.05. The scores of those using their smartphones for 3–4 hours each day were not significantly different from those using them for 4–5 hours per day (P = 0.33). However, there were significantly large effects (ƞ² = 0.2) in those who used smartphones for less than 3 hours versus 3–4, 4–5, and more than 5 hours (P < 0.001), in ascending order of score values.

DISCUSSION

SAS-SV was developed for South Korean adolescents by Kwon et al.[25] to identify behaviors that indicated PSU. SAS-SV has emerged as a reliable and valid brief measure to assess PSU among adolescents and adults across countries.[17] Although SAS-SV has been used in Indian studies, it is yet to be validated for use in this population.[2,14] Therefore, the current study aimed to validate SAS-SV for the Indian adult population.

After the expert review, the items were reworded into sentences that had easier comprehensibility, fewer colloquialisms, and fuller sentences. The beginnings of all the sentences were reworded to include first-person language. Descriptors, such as “Can’t stand it,” were changed to more concrete descriptors, such as “tolerate.”

The CFA did not show good fit indexes as in De Pasquale and colleagues’ study,[33] for instance, CFI = 0.92, and SRMR = 0.06. Hence, an EFA method was used similarly to other SAS-SV adaptation studies.[31,34,35] This procedure showed good validity and adequacy of the instrument, and its single factor explained 44% of the variance. This variance found in the current validation was higher than that in the Arabic (42.4%) and Brazilian (39.2%)[31,35] and lower than the Spanish (49%) and Belgian (54%) versions of SAS-SV.[34]

The Indian validation of SAS-SV had Cronbach’s alpha of 0.85, similar to the original scale’s 0.91[25] and nearer to the 0.84 of the U.S. validation.[26] Similarly, the current values on internal consistency are similar to Turkey (α = 0.88),[36] Spain and Belgium (α = 0.88 and α = 0.90, respectively),[34] Italy (α = 0.79),[33] Morocco (α = 0.87),[35] Hong Kong (α = 0.84),[37] Brazil (α = 0.81),[31] Serbia (α = 0.89),[38] and Pakistan (α =0.81).[9] A satisfactory test-retest reliability with an ICC value of 0.81 was also obtained. Similar values were found in Brazil (ICC = 0.85),[39] Hong Kong (ICC = 0.76),[37] and Serbia (ICC = 0.94).[38] The findings from the current study support the use of the SAS-SV India validated scale as a good measure of PSU assessment.

The cut-off values for the severity of PSU proposed in the original study were limited to the Korean adolescent sample, and using the same cut-off values for the current study of Indian young adults would be incorrect and possibly result in misdiagnosis. The current study avoided considering SAS-SV as a diagnostic tool, especially considering the absence of smartphone addiction in diagnostic manuals. It was rather aimed at guiding clinicians to identify PSU severity and track treatment progress. Therefore, the current study adopted the percentile distribution method to understand the scores. The distribution was divided into two levels: gender and age groups. The rationale for arriving at the multilevel distribution was derived from the analysis of the differences in the means of these age categories and to provide more nuance to the test administrator. The current study considered the 75^th percentile as an indicator of a significantly high PSU. Therefore, among the total sample, a score of 36 and above was considered a high PSU. This value was near the cut-offs found in different countries.[25,26,40]

There were no significant score differences based on gender, and these observations agree with the Spanish and Belgian study[34] and the U.S. study.[26] Higher scores on PSU were observed in younger participants, a finding that was consistent with the literature.[31] Marital status did not have any significant score differences, similar to the findings in Spain and Belgium[34] and the US.[26] Less than 3 hours of smartphone usage had lesser scores than those with over 5 hours of usage, an observation consistent with previous studies.[27] The scores were higher for those with an education up to 12^th standard and under graduation versus post-graduation and above. These findings are consistent with the original study on the adult version of the scale conducted in Korea[41] and in Tehran.[42] Studies on adolescents have reported the highest PSU trends, and as reported in the current study, this trend was reduced significantly by an increase in age and educational qualifications.[11,25,27,43,44]

Strengths of the study: As SAS-SV was widely used unstandardized in the Indian context, the current study helped in its validation for the Indian population. It had a robust expert validation process that quantitatively and qualitatively evaluated the items for the Indian context. The study had a large sample size across various regions of India. The sociodemographic analyses nuanced the understanding of PSU in the Indian context. Employing percentile distribution rather than cut-offs helped retain the construct of smartphone addiction open to future research instead of limiting it to diagnostic categorization. The percentile ranges would also help in tracking treatment progress due to a finer understanding of PSU changes rather than a dichotomous observation of the presence or absence of smartphone addiction.

Limitations of the study: The study lacked a concurrent validity index. Convenience sampling was employed to collect data. The scale was validated in the English language and is thus limited to English-speaking Indian citizens.

CONCLUSION

This study is the first to validate SAS-SV for use with Indian adults, despite its frequent use with this population in India. SAS-SV has retained its psychometric properties when examined cross-culturally and is valid and reliable for assessing PSU in Indian adults. SAS-SV would be beneficial in the early detection and tracking of PSU. In the future, studies can focus on translating SAS-SV to regional Indian languages and validation on the child and adolescent population. The final version of the scale is presented in Table 7.

Table 7.

Smartphone Addiction Scale – Short Version – India (SAS-SV-India) Below is a collection of statements about your everyday smartphone use. Using the 1–6 scale below, please indicate how frequently or infrequently you currently have each experience

S.NO	Smartphone Addiction Scale - Short Version - India- Items	Strongly disagree	Disagree	Slightly disagree	Slightly agree	Agree	Strongly agree
1.	I miss planned work due to smartphone use.	1	2	3	4	5	6
2.	I have difficulties concentrating in class, while doing assignments, or while working due to smartphone use.	1	2	3	4	5	6
3.	I experience discomfort in my wrists, thumbs, eyes, or at the back of the neck due to smartphone use.	1	2	3	4	5	6
4.	I will not be able to tolerate not having a smartphone.	1	2	3	4	5	6
5.	I feel impatient or irritable when I am not holding my smartphone.	1	2	3	4	5	6
6.	I keep thinking of my smartphone even when I am not using it.	1	2	3	4	5	6
7.	I may not be able to give up my smartphone even though my daily life is affected by it.	1	2	3	4	5	6
8.	I constantly check my smartphone so as not to miss any conversations on social media (WhatsApp, Instagram, Facebook, Reddit, Twitter, etc.)	1	2	3	4	5	6
9.	I use my smartphone longer than I intended or planned.	1	2	3	4	5	6
10.	My family or friends have told me that I use my smartphone too much.	1	2	3	4	5	6

Total Score = , Percentile =

Authors’ contributions

Conception and design of the study: GF, MKS, NA. Acquisition and interpretation of data GF, MKS, NA BB, KS. Drafting the article: GF, MKS, NA. Revising it critically NA. Final approval of the manuscript version to be submitted: GF, MKS, NA BB, KS.

Ethical statement

The study has been approved by the Institute Ethics Committee, National Institute of Mental Health and Neuro Sciences (NIMHANS), Institute of National Importance (INI), Bangalore, Karnataka, India.

Conflicts of interest

There are no conflicts of interest.

Data availability statement

All data is reported in this manuscript.

Funding Statement

Nil.