Data Resource Profile: The National Cancer Institute’s Health Information National Trends Survey (HINTS)

Data resource basics

The National Cancer Institute’s (NCI) Health Information National Trends Survey (HINTS) was conceived in 1997 during a multidisciplinary conference focused on risk communication with attenders representing the fields of psychology, health behaviour and education, public health, clinical medicine and health journalism. The key recommendation born of this conference encouraged NCI to develop a communication-specific population survey to track trends in US adults' access to, need for and use of health and cancer information. Heeding the call for development of a national communication survey, NCI developed a nationally representative survey to assess trends in cancer-related communication, health information-seeking and cancer-related knowledge, attitudes and behaviour.

HINTS is a cross-sectional, nationally representative survey of the US non-institutionalized adult population (aged 18 years and older) which collects data on health-related information and health-related knowledge, attitudes and behaviour.^1,2 HINTS was first fielded in 2002–03² and the general population survey has been administered five times over a 15-year period, with HINTS 4 and 5 including multiple annual cross-sectional data collection cycles. The resulting data (n = 37 365) can be tracked for trends over time or, if there are no trends anticipated a priori, can be aggregated into a larger sample for further analysis. Table 1 summarizes survey design and implementation details for each completed survey administration, including survey field period, survey mode, total sample size, response rate, and number of cancer patients and survivors.

Table 1.

Health Information National Trends Survey (HINTS) survey design and implementation characteristics (2003–17)

	HINTS 1	HINTS 2	HINTS 3	HINTS 4				HINTS 5
				Cycle 1	Cycle 2	Cycle 3	Cycle 4	Cycle 1
Field period initiated	2002	2005	2008	2011	2012	2013	2014	2017
Mode	RDD	RDD	Mail & RDD	Mail	Mail	Mail	Mail	Mail
Total no. of respondents	6369	5586	Mail: 3582 RDD: 4092	3959	3630	3185	3677	3285
Response rate	33.0%	24.0%	Mail: 40.0% RDD: 24.2%	36.7%	40.0%	35.2%	34.4%	32.4%
No. cancer patients/survivors	763	873	1001	563	464	459	542	504

Data collected

HINTS 1 was administered in 2002–03 as a random digit dial (RDD) computer-assisted telephone interview to a representative sample of US households drawn from all telephone exchanges in the US. One adult aged 18 years or older within each household was selected for the extended interview during a household screening. Interviews were conducted in English or Spanish, depending on respondent preference. Further details about the sample and sampling design are published elsewhere.²

HINTS 2 was administered in 2005 as an RDD computer-assisted telephone interview to a representative sample of U.S. households drawn from all telephone exchanges in the U.S. One adult from each household was selected for interview, which was conducted in English or Spanish according to respondent preference. Further details about sampling design are published elsewhere.³

HINTS 3 was fielded in 2008 using a mixed-mode, dual-frame design. One sample frame was a list-assisted RDD computer-assisted telephone interview, wherein one adult from each household was selected for an interview. Interviews were conducted in English or Spanish, depending on respondent preference. The second sample frame was a comprehensive national listing of addresses available from the United States Postal Service. These households were administered a mailed survey. In the mail sample, all adults in the household at each sampled address were asked to complete a questionnaire. Thus, the mail sample was a stratified cluster sample, in which the household was the cluster. Further details on the HINTS 3 survey design and operations are published elsewhere.⁴

The HINTS 4 administration included four cross-sectional mail-mode data collection cycles over 4 years starting in 2011 and concluding in 2014 (Table 1). HINTS 4 Cycles 1–4 were administered as mailed questionnaires using a sampling frame of addresses provided by Marketing Systems Group (MSG). The protocol for mailing the questionnaires involved an initial mailing of the questionnaire, followed by a reminder postcard, and up to two additional mailings of the questionnaire as needed for non-responding households. Most households received one survey per mailing (in English), whereas households that were potentially Spanish-speaking received two surveys per mailing (one in English and one in Spanish). In the second stage of sampling, one adult from each sampled household was selected for participation. Further details on the survey design and operations for the HINTS 4 data collection cycles have been previously described.^5–8

HINTS 5 includes four cross-sectional data collection cycles over 4 years, starting in 2017 and scheduled to end in 2020. The first of the HINTS 5 data collection cycles (HINTS 5, Cycle 1) was conducted in 2017. HINTS 5, Cycle 1 was administered as a single-mode mailed survey using a sampling frame of addresses provided by MSG, following the same protocol as HINTS 4. Further details on the HINTS 5, Cycle 1 data collection have been previously published.⁹

Three additional topic-specific HINTS modules, not described in this data resource profile, were fielded in 2009 (HINTS Puerto Rico), 2015 (HINTS-FDA Cycle 1) and 2017 (HINTS-FDA Cycle 2).

Data quality

To ensure data quality before data collection, each HINTS administration has included cognitive testing for each HINTS instrument, with testing particularly focused on new questions. The goal of cognitive testing is to provide valid measures of the constructs of interest with a minimum of response error.^10–12 Testing was conducted on both paper and RDD surveys.

After data collection, data quality efforts for the RDD surveys include direct data entry. The computer-assisted telephone interview (CATI) program ensures that proper skip patterns are followed and constrains data entry to valid values for each survey item. The CATI program also allows for range and edit checks on the entry of numerical responses to questions such as age, length of time since last search for health information, servings of fruits or vegetables consumed daily, height and weight and other numerical response questions.

For mailed surveys, post-data collection quality control checks are conducted on the scanned data and electronic images of the survey. Quality assurance staff compare the hard-copy questionnaire with the data captured in the database item-for-item and the images stored in the repository page-for-page, to ensure that all items are correctly captured. Scanned data are validated according to HINTS specifications. Violations of validation rules (such as marking more than one choice box in a mark-only-one question) are flagged and reviewed. Additionally, quality assurance staff closely review frequencies and cross tabulations of the HINTS raw data to identify outliers and to verify open-ended items.

Survey constructs and measures

Each HINTS administration includes a core set of items. Table 2 summarizes the key constructs represented in the survey core of HINTS and describes the nature of the associated survey items. Core constructs include health information-seeking, cancer prevention and screening, cancer-related knowledge and behaviour; cancer risk perceptions; health care use and access; and technology utilization and access. Additional constructs or items have been included in specific HINTS administrations to capture timely or emerging phenomena or to glean greater detail on core constructs. Table 3 summarizes the sociodemographic characteristics of the total sample and weighted estimates for each of the survey administrations. Estimates are weighted to correct for non-response bias and to be representative of the US population; population distributions therefore reflect those of the US population with regard to age, sex, race and ethnicity.

Table 2.

Core constructs measured in the Health Information National Trends Survey (HINTS)

Construct	Measures
Sociodemographics	Age, sex, race, ethnicity, income, home ownership status, financial strain, health insurance coverage, education, marital status, employment status, country of origin (US vs other), health status, cancer history
Health information-seeking	Ever sought health information, health information sources, trust in health information sources, confidence in health information-seeking, information-seeking experiences, internet use for health information
Cancer prevention and screening knowledge and behaviour	Colorectal, breast, and cervical cancer screening, HPV vaccination
Cancer-related behaviour	Tobacco use, sun safety, physical activity, diet
Cancer risk perceptions	Confusion, fatalism, health beliefs, perceived risk of developing cancer
Health care use and access	Usual source of care, cost barrier to care, patient-provider communication
Technology use and access	Internet access through dial-up, broadband, cellular network, wireless network; use of internet for health-related reasons; ownership of tablet computers, smartphones, basic cellphones; use of health-related apps; use of social media for health-related reasons

Table 3.

Sample size and weighted estimates for HINTS respondents by sociodemographic characteristics

	HINTS 1		HINTS 2		HINTS 3		HINTS 4								HINTS 5		Combined
	2003		2005		2008		Cycle 1 (2011)		Cycle 2 (2012)		Cycle 3 (2013)		Cycle 4 (2014)		Cycle 1 (2017)		Total
	N	%	N	%	N	%	N	%	N	%	N	%	N	%	N	%	N	%
Total	6369	11.4	5586	11.7	7674	12.2	3959	12.7	3630	12.9	3185	12.7	3677	12.9	3285	13.4	37 365	100.0
Sex
Female	3848	51.9	3657	51.9	4696	51.4	2304	51.5	2172	51.4	1906	51.6	2184	51.7	1914	51.1	22 681	51.6
Male	2521	48.1	1929	48.1	2969	48.6	1552	48.5	1390	48.6	1197	48.4	1424	48.3	1303	48.9	14 285	48.4
Age
18-34	1656	31.2	1037	31.0	1113	30.8	582	30.3	529	30.5	426	27.2	467	30.8	367	21.9	6177	29.1
35-49	1961	31.1	1494	30.1	1831	29.6	932	27.3	845	26.4	712	30.5	743	26.7	655	28.7	9173	28.8
50-64	1492	21.5	1522	22.8	2451	23.1	1339	25.2	1168	25.6	1070	25.1	1220	25.1	1063	30.1	11 325	24.9
65-74	694	9.7	812	9.3	1189	8.4	583	9.2	555	9.4	514	9.4	637	9.6	676	11.1	5660	9.5
>=75	548	6.6	707	6.7	1010	8.1	455	8.0	414	8.0	360	7.9	428	7.9	385	8.2	4307	7.7
Race/ethnicity
Hispanic	764	11.7	496	13.0	622	12.8	461	14.5	511	15.0	511	15.4	540	15.1	427	15.7	4332	14.2
NH White	4276	71.8	4103	69.9	5445	69.3	2431	66.8	2043	67.1	1584	66.9	1960	66.6	1868	65.7	23 710	67.9
NH Black	716	10.5	438	10.0	687	11.5	576	11.4	496	10.9	421	10.5	534	11.3	409	10.3	4277	10.8
NH Other	312	6.0	299	7.1	424	6.4	271	7.4	208	7.1	209	7.2	239	6.9	249	8.3	2211	7.1
Education
Less than high school	747	16.9	687	14.5	683	13.9	391	12.8	329	13.5	297	9.7	308	11.6	217	8.7	3659	12.6
High school graduate	1828	32.0	1447	29.9	1804	26.6	785	23.1	775	20.3	699	24.4	670	18.2	616	22.9	8624	24.5
Some college	1637	26.8	1545	32.2	2192	34.8	1167	31.1	1057	37.6	933	32.7	1090	30.0	942	32.8	10 563	32.3
College graduate	1927	24.3	1696	23.5	2637	24.7	1531	32.9	1380	28.6	1167	33.2	1458	40.1	1406	35.6	13 202	30.6
Income
<$20, 000	1 111	19.2	899	17.7	1142	19.8	829	24.3	740	21.9	680	20.2	774	19.4	559	17.4	6734	20.0
$20, 000 to <$35, 000	1295	22.5	868	17.7	1056	16.7	584	17.2	501	14.9	418	14.5	489	12.7	423	12.2	5634	15.9
$35, 000 to <$50, 000	958	17.6	652	14.2	873	14.0	520	12.6	459	15.5	394	14.7	482	14.8	386	14.9	4724	14.7
$50, 000 to <$75, 000	955	17.7	924	22.0	1203	19.1	594	16.9	524	16.8	446	17.7	550	17.2	530	19.1	5726	18.2
$75, 000+	1214	23.1	1150	28.4	2041	30.4	1031	29.0	926	31.0	801	33.0	979	35.9	1064	36.3	9206	31.1
Metro/non-metro county
Metro	5174	80.9	4352	80.0	6192	82.2	3321	84.2	3087	83.7	2709	82.1	3157	83.6	2812	84.4	30 804	82.7
Non-metro	1195	19.1	1234	20.0	1482	17.8	638	15.8	543	16.3	476	17.9	520	16.4	473	15.6	6561	17.3

Regulatory and ethical considerations

Each HINTS administration has been approved through expedited review by the Westat Institutional Review Board, and subsequently deemed exempt by the U.S. National Institutes of Health Office of Human Subjects Research Protections.

Data resource use

HINTS data are used by researchers to explore use of different communication channels to obtain health information among the US adult population; these data are also used to assess public knowledge and attitudes about health-relevant topics. HINTS data are used by programme planners to identify health information and communication facilitators and barriers within and across populations, and to inform the development of effective health communication strategies. Social scientists use HINTS data to test and refine their theories of health communication in the information age and to guide recommendations for theory-driven interventions aimed at improving population health.

HINTS data have been used to pursue a wide variety of research questions. To date, there have been more than 400 peer-reviewed articles, published in more than 160 journals, which have used HINTS data, and an edited dedicated book comprising the HINTS knowledge base.¹³ As an illustrative example, two special issues of the Journal of Health Communication have featured HINTS articles, following from HINTS research presented at the national HINTS Data Users Conferences. The first special issue, The Health Information National Trends Survey (HINTS): Research from the Baseline, was published in 2006, featuring data from the inaugural HINTS data collection.¹ Studies published in this special issue covered a range of topics including cancer knowledge,^14,15 cancer cognition and risk perceptions,^16–20 and cancer information-seeking and communication.^21–25 The second special issue, Partners in Progress: Informing the Practice of Health Communication through National Surveillance, was published in 2010.²⁶ This special issue featured analysis of data from the first three administrations of HINTS and included studies on the following topics: health communication surveillance methodology,^27,28 health communication and information-related disparities,^29–37 patient-provider communication,^38–40 and use of the internet and health communication technology.^33,41–43

Highly cited articles from these special issues offer examples of the specific research topics pursued through use of HINTS data. Viswanath and colleagues examined the relationship between publicity and knowledge gaps using two cancer topics with varied levels of publicity: knowledge about tobacco and sun exposure.¹⁴ Results indicated that education and income were associated with awareness of the smoking and cancer link despite heavy media attention, and having at least a high school education was associated with knowledge of the sun exposure and cancer link.¹⁴ Dillard and colleagues assessed whether perceived risk of developing lung cancer was associated with acceptance of smoking-related myths and beliefs.¹⁹ Those whose perceived that risk was less than their objective risk (unrealistic optimists) were more likely to report: that there is no risk of developing lung cancer among those who smoke only a few years; and that developing lung cancer is determined by genetic factors. Unrealistic optimists were less likely to report an intention to quit smoking.¹⁹ Koch-Weser and colleagues examined sociodemographic factors associated with use of the internet for health information-seeking and found that those who used the internet to seek health information were younger, had higher education and income, preferred numbers to words to describe risk and were more likely to indicate that it is important to get personal medical information electronically.⁴²

A complete bibliography of HINTS publications is maintained on the HINTS website: [https://hints.cancer.gov/publications-reports/hints-publications.aspx]. A variety of other HINTS reports are also available: [https://hints.cancer.gov/publications-reports/nci-reports.aspx].

A series of HINTS Briefs have been developed to provide population-level estimates for specific survey questions and to summarize noteworthy research findings by demographic characteristics: [https://hints.cancer.gov/publications-reports/hints-briefs.aspx]. HINTS Briefs are targeted toward results users who want to use HINTS data to inform their applied work. Many HINTS Briefs also summarize research findings from recent peer-reviewed journal articles using HINTS data. The following are some of the topics summarized in HINTS Briefs: Considerations for Developing Effective Health Communication Strategies; Preventing Cancer through Increased Human Papillomavirus (HPV) Vaccine Uptake; and Developing an Electronic Health Information System for High-Quality Cancer Care.

Cancer patients and survivors

HINTS data have been extensively used to characterize the experiences of cancer patients and survivors. We conducted a review of the literature to identify published studies using HINTS data focused on cancer patients and survivors. We searched MEDLINE and EMBASE from 1 January 2003 to 15 May 15 2018, using the following search terms: health information national trend* or HINTS as text phrases, AND survivor* or cancer* OR explode neoplasms [MeSH]. This search returned 229 abstracts. All abstracts were reviewed to identify those meeting the following conditions: (i) the study used data from U.S. national collection of the Health Information National Trends Survey (HINTS); and (ii) the study focused on cancer survivors/patients or compared cancer survivors or patients with other populations. Studies that used HINTS items or a HINTS instrument to collect data in a special (non-national) population were excluded, and studies that used cancer history as a control variable (i.e. not a variable of specific interest) were also excluded. A total of 35 articles met the inclusion criteria and were reviewed. The following themes emerged in the focus of the articles reviewed: information seeking trends, experiences and sources^44–51; patient-centred communication and clinical care^52–55; use of internet and mobile technology in health^56–62; health-related behaviour^63–71; cancer cognition^72–74; and health status and health outcomes.^75–77

Table 4 summarizes the sociodemographic, cancer diagnosis and treatment status characteristics of the cross-sectional cohorts of cancer patients and survivors for each relevant HINTS administration. The table summarizes the total number of cancer patients per each survey administration, describes the sociodemographic characteristics thereof and indicates counts and percentages for specific cancer types. Across the survey years, the most frequent cancer types were breast, colon, cervical, skin, prostate, melanoma and endometrial cancer. In each survey year, most cancer patients reported receiving treatment for their cancer (range: 81.5–91.8%), with the majority of respondents reporting at least 2 years since diagnosis. These data offer a rich resource for examination of cancer patients’ and survivors’ cancer-related knowledge, attitudes and behaviours, as well as their information-seeking experiences and needs.

Table 4.

Sample size and weighted estimates for HINTS respondents with a personal history of cancer by sociodemographic characteristics and cancer-related characteristics

	HINTS 1		HINTS 2		HINTS 3		HINTS 4								HINTS 5		Combined
	2003		2005		2008		Cycle 1 (2011)		Cycle 2 (2012)		Cycle 3 (2013)		Cycle 4 (2014)		Cycle 1 (2017)		Total
	N	%	N	%	N	%	N	%	N	%	N	%	N	%	N	%	N	%
Have you ever been diagnosed as having cancer?
Yes	763	10.77	873	11.31	1, 001	7.25	563	8.53	464	8.19	459	8.08	542	8.52	504	8.64	5, 169	8.88
Sex
Female	532	64.57	596	56.09	590	56.65	317	57.83	258	54.77	280	59.96	319	62.41	303	59.68	3195	59.08
Male	231	35.43	277	43.91	411	43.35	229	42.17	200	45.23	168	40.04	205	37.59	187	40.32	1908	40.92
Age
18-34	57	8.78	29	5.09	18	2.54	14	3.08	15	5.80	12	4.66	9	3.78	4	1.71	158	4.56
35-49	139	19.71	117	17.55	101	14.89	46	10.08	45	11.78	40	15.37	48	15.07	30	10.50	566	14.53
50-64	229	30.44	246	32.18	323	34.10	202	35.00	138	31.95	137	29.64	164	34.82	171	37.16	1610	33.10
65-74	164	21.29	233	22.78	278	22.36	144	24.98	126	24.59	141	25.85	146	22.30	153	24.95	1385	23.59
>=75	173	19.78	246	22.40	270	26.12	145	26.86	135	25.87	120	24.49	146	24.03	125	25.67	1360	24.21
Race/ethnicity
Hispanic	38	4.69	33	5.12	40	4.96	27	4.72	42	8.92	37	5.77	51	11.24	36	8.51	304	6.67
NH White	613	82.70	730	84.58	844	84.37	411	82.61	320	80.42	285	84.43	337	80.38	340	76.18	3880	81.93
NH Black	51	7.51	37	5.03	54	8.18	51	6.36	39	6.52	36	7.08	50	6.23	51	11.36	369	7.24
NH Other	31	5.09	38	5.27	35	2.50	32	6.30	18	4.13	16	2.72	22	2.15	22	3.94	214	4.15
Education
Less than high school	86	17.07	105	14.19	100	15.97	56	13.81	42	16.11	46	11.54	48	15.70	30	6.86	513	13.86
High school graduate	246	37.02	233	31.48	242	26.70	140	26.38	106	19.02	106	25.12	94	16.00	117	27.53	1284	26.51
Some college	185	22.85	266	30.73	268	28.77	161	26.90	151	39.75	132	30.24	164	31.95	147	34.72	1474	30.66
College graduate	228	23.06	239	23.60	384	28.55	193	32.91	160	25.12	162	33.10	209	36.36	193	30.89	1768	28.97
Income
<$20, 000	150	22.12	172	22.02	160	21.11	128	24.53	84	18.84	96	17.55	114	18.52	90	18.40	994	20.42
$20, 000 to <$35, 000	177	25.63	143	20.58	161	18.61	88	16.91	64	15.50	46	11.82	79	13.95	74	13.24	832	17.14
$35, 000 to <$50, 000	106	18.78	95	14.29	134	17.31	81	18.00	59	15.67	61	14.22	70	14.98	51	9.75	657	15.28
$50, 000 to <$75, 000	100	14.39	127	19.89	146	16.61	74	14.69	64	18.46	73	20.79	73	13.68	82	22.21	739	17.61
$75, 000+	117	19.07	141	23.22	241	26.36	121	25.87	111	31.54	109	35.63	137	38.88	154	36.40	1131	29.54
Metro/non-metro county
Metro	613	79.54	676	78.30	798	79.33	459	82.23	376	74.03	389	83.75	461	80.16	426	81.99	4198	79.89
Non-metro	150	20.46	197	21.70	203	20.67	104	17.77	88	25.97	70	16.25	81	19.84	78	18.01	971	20.11
Derived variable to categorize responses on cancer typea
Bladder cancer only	10	1.38	13	1.78	17	1.54	4	0.74	8	1.46	6	1.52	11	1.30	5	0.76	74	1.31
Bone cancer only	1	0.23	2	0.19	5	0.79	1	0.09	0	.	0	.	2	0.12	1	0.01	12	0.16
Breast cancer only	108	13.11	164	14.25	139	13.02	77	10.93	65	12.54	72	17.13	87	15.36	83	14.72	795	13.89
Cervical cancer only	90	11.32	60	7.74	66	8.98	37	7.46	33	8.39	30	5.46	30	7.10	29	8.67	375	8.18
Colon cancer only	41	5.54	37	4.44	50	5.02	30	5.10	15	2.61	17	3.42	20	3.31	23	5.08	233	4.35
Endometrial cancer only	36	4.88	35	2.79	31	2.64	13	3.26	10	2.87	11	1.92	15	4.11	10	1.58	161	3.04
Head/neck cancer only	8	0.81	5	0.77	5	0.75	7	1.35	3	0.72	4	0.92	10	1.85	6	1.47	48	1.08
Hodgkins only			6	1.03	4	0.34	4	0.57	8	4.17	2	0.42	7	1.13	4	3.16	35	1.36
Renal cancer only	11	1.69	12	1.29	10	0.71	5	0.78	5	1.21	10	2.03	9	2.13	4	0.51	66	1.30
Leukaemia only	4	0.51	3	0.49	13	1.36	8	1.23	3	0.59	5	0.72	7	2.49	7	1.44	50	1.08
Liver cancer only			2	0.22	1	0.30	2	0.08	0	.	0	.	3	0.78	1	0.49	9	0.23
Lung cancer only	13	1.92	9	1.77	16	1.85	6	1.15	10	2.66	6	1.74	8	1.10	8	2.07	76	1.78
Melanoma only	52	7.66	68	8.78	67	7.75	26	5.21	19	4.39	20	5.97	27	4.99	23	4.58	302	6.22
Non-Hodgkin only			9	1.42	8	1.07	6	1.22	7	1.30	6	1.95	4	0.35	9	2.08	49	1.17
Oral cancer only	6	0.88	2	0.25	3	0.20	3	0.70	4	0.95	1	0.13	1	0.17	0	.	20	0.42
Ovarian cancer only	21	2.69	21	3.46	16	1.46	13	2.64	6	0.63	4	0.60	4	0.52	2	0.05	87	1.59
Pancreatic cancer only	2	0.23	1	0.07	3	0.24	0	.	1	0.04	3	0.17	0	.	2	0.35	12	0.14
Pharyngeal cancer only	5	0.60	6	1.60	6	1.35	1	0.05	1	0.09	1	0.11	2	0.67	0	.	22	0.57
Prostate cancer only	61	9.62	71	9.72	87	8.48	53	9.97	51	10.76	54	9.47	59	8.33	42	6.20	478	9.08
Rectal cancer only	2	0.13	2	0.09	2	0.15	3	0.18	0	.	0	.	2	0.21	4	0.78	15	0.20
Skin cancer only	138	16.30	162	17.88	239	22.44	116	23.08	92	20.05	107	25.40	117	25.02	124	24.95	1095	21.67
Stomach cancer only	5	1.45	9	1.65	3	0.35	0	.	3	0.50	2	0.17	1	0.01	2	0.71	25	0.66
More than one cancer checked	60	7.54	115	11.32	140	14.20	107	18.44	82	18.21	66	14.17	65	13.93	70	13.63	705	13.72
Other cancer only	62	8.71	57	7.01	50	4.99	32	5.76	26	5.86	24	6.57	28	5.01	34	6.72	313	6.42
Lymphoma only (HINTS 1)	19	2.81															19	0.39
Did you ever receive any treatment for your cancer?
Yes			721	81.45	862	86.92			417	91.01			477	89.13	457	91.76	2934	87.83
No			151	18.55	120	13.08			44	8.99			55	10.87	40	8.24	410	12.17
How long ago were you diagnosed with cancer? (derived)
Less than 1 year since diagnosis	115	16.62	119	15.95	99	10.90	58	9.88	56	10.00	58	14.49	57	12.43	61	15.22	623	13.43
2-5 years	171	20.32	231	31.11	241	26.61	127	23.42	111	28.42	117	31.10	113	19.92	96	17.66	1207	24.82
6-10 years	164	23.43	160	16.06	184	17.95	101	20.24	84	20.31	88	17.73	94	21.14	95	20.73	970	19.71
11+ years	305	39.64	351	36.88	442	44.54	247	46.46	190	41.27	170	36.69	221	46.51	212	46.39	2138	42.04

^aHINTS 1 had different response options that did not have liver, Hodgkin and non-Hodgkin cancers. Hodgkin and non-Hodgkin cancers were probably combined under lymphoma.

Strengths and weaknesses

HINTS is unique among national data resources in its focus on health communication and health information. The HINTS programme offers a resource for investigators from diverse disciplines and gives access to data that speak to population use of information and communication resources during a time of unprecedented change in the information and communication landscape. As described above and in Table 4, HINTS also provides a rich cohort of cancer patient and survivor data for secondary analysis.

Since its inception, the HINTS programme has invested heavily in efforts to ensure that the data are readily and easily accessible and usable for data users and results users. The HINTS website is rich with tools to enable data access and to support data use (see Data resource access section). A variety of materials have also been developed for results users, including an online data display tool, and HINTS Briefs, which summarize key results from HINTS publications and offer insights into potential applications in public health and clinical practice.

HINTS data are cross-sectional; as such, inferences about causality are generally not appropriate. Similar to other national health surveys, response rates for HINTS have declined over time.^78,79 Lower response rates may introduce bias in the data, but efforts were made in each HINTS administration to reduce potential for bias through sampling and weighting procedures.⁴ Additionally, methodological research suggests that the negative impact of declining response rates on data quality may not be as dramatic as previously assumed.^78,80,81 The HINTS programme team recently conducted a rigorous non-response bias analysis of data from HINTS 4 (Cycles 1 and 3) to characterize the potential impact of non-response.⁸² Findings from this study revealed that many of the demographic influences on non-response (e.g. age, socioeconomic status) can be compensated for through application of standard weighting procedures.⁸² More specific results of this non-response analysis involving comparison of response rates among population subgroups, comparison with national benchmarks, and level-of-effort analysis have been previously published.⁸²

National surveys are usually constrained to measuring constructs with only one or two survey items, to reduce respondent burden. Therefore, the number of items available for measuring complex attitudinal and behavioural constructs is often limited. Although use of single items for measurement of social and behavioural constructs is not ideal, this approach is common in large-scale and national survey research.^17,83–85 When compared with validated multi-item scales, single-item measures can have similar test–retest reliability and construct validity^86,87; however, single items are less reliable than scales and may attenuate observed associations.⁸⁸ Constraints on survey length may also lead to changes in the survey content over time, thereby limiting the temporal trends and comparisons that may be tracked over time.

While the relatively small sample size for HINTS does not support the calculation of reliable estimates at the state level, innovative geographical information system strategies have been employed to calculate regional geographical estimates,^89–91 and supplemental funding has been granted to certain NCI-designated comprehensive cancer centres to support collection of HINTS data at the local level.⁹²

Data resource access

The HINTS programme has striven from the beginning to enable users to use the HINTS data as easily as possible. HINTS data are in the public domain, and public use datasets are available for download as SAS, STATA and SPSS files from [https://hints.cancer.gov]. Each dataset is bundled with supporting documents including analytics recommendations (including example statistical software code), history document, codebook, methodology report (including details about sampling and the creation of survey weights) and annotated survey instruments. Codebooks for each dataset are also available online, and all of the instruments included in the data download bundle are annotated with variable names and allowable codes. Additionally, several resources been published describing how to use HINTS data: [https://hints.cancer.gov/meetings-trainings/how-to-hints-webinar.aspx]. Where possible, HINTS adheres to the F.A.I.R. (Findable, Accessible, Interoperable, and Reusable) principles.⁹³

Profile in a nutshell

• HINTS is as a cross-sectional national survey of non-institutionalized adults, developed to track trends in cancer-related communication, health information-seeking and cancer-related knowledge, attitudes and behaviour in the US population. It is the only national population-based survey that collects information on the US public’s need for, access to and experience with cancer-related information.

• HINTS was first fielded in 2002–03, and has been administered five times over approximately a 15-year period, with HINTS 4 and 5 including multiple data collection cycles (n = 37 365).

• HINTS was initially administered as a random digit dial (RDD) computer-assisted telephone interview to a representative sample of households drawn from all telephone exchanges in the US. In 2008, HINTS was fielded using a mixed-mode (RDD telephone interview and mailed questionnaire), dual-frame (all telephone exchanges in the US and a comprehensive national listing of United States Postal Service addresses) format. Each HINTS administration since 2008 has been conducted as a mailed survey using an address-based sampling frame.

• Each HINTS instrument includes a core set of items to assess: communication technology access and use, health and cancer information-seeking, cancer-related knowledge and behaviour, cancer risk perceptions and health care access.

• HINTS public use datasets are available for download: [https://hints.cancer.gov].

Funding

This work was supported by the National Cancer Institute via HHSN261201800002B.

Conflict of interest: None declared.